Escolar Documentos
Profissional Documentos
Cultura Documentos
Contents
I. Overview 2
II. Introduction to Cytoplasmic One-Carbon Metabolism 4
A. Enzymes that generate one-carbon units 5
B. Folate-interconverting enzymes 13
C. Biosynthetic enzymes 18
D. Folate-binding proteins 24
III. Introduction to Mitochondrial One-Carbon Metabolism 24
A. Enzymes that generate one-carbon units 25
B. Folate-interconverting enzymes 27
C. Biosynthetic enzymes 28
IV. Nuclear Folate-Mediated One-Carbon Metabolism 28
Acknowledgments 29
References 29
Abstract
Tetrahydrofolate (THF) polyglutamates are a family of cofactors that carry and
chemically activate one-carbon units for biosynthesis. THF-mediated one-
carbon metabolism is a metabolic network of interdependent biosynthetic path-
ways that is compartmentalized in the cytoplasm, mitochondria, and nucleus.
One-carbon metabolism in the cytoplasm is required for the synthesis of
purines and thymidylate and the remethylation of homocysteine to methionine.
One-carbon metabolism in the mitochondria is required for the synthesis of
formylated methionyl-tRNA; the catabolism of choline, purines, and histidine;
and the interconversion of serine and glycine. Mitochondria are also the primary
source of one-carbon units for cytoplasmic metabolism. Increasing evidence
indicates that folate-dependent de novo thymidylate biosynthesis occurs in
the nucleus of certain cell types. Disruption of folate-mediated one-carbon
metabolismis associated with many pathologies and developmental anomalies,
Vitamins and Hormones, Volume 79
#
2008 Elsevier Inc.
ISSN 0083-6729, DOI: 10.1016/S0083-6729(08)00401-9 All rights reserved.
* Graduate Field of Biochemistry, Molecular and Cellular Biology, Cornell University, Ithaca,
New York 14853
{
Division of Nutritional Sciences, Cornell University, Ithaca, New York 14853
1
yet the biochemical mechanisms and causal metabolic pathways responsible
for the initiation and/or progression of folate-associated pathologies have yet
to be established. This chapter focuses on our current understanding of mam-
malian folate-mediated one-carbon metabolism, its cellular compartmentation,
and knowledge gaps that limit our understanding of one-carbon metabolism
and its regulation. 2008 Elsevier Inc.
I. Overview
The reduced tetrahydrofolates (THFs) serve as a family of enzyme
cofactors that chemically activate and carry one-carbon units on the N5 and/
or N10 of THF at the oxidation level of formate (e.g., 10-formylTHF),
formaldehyde (e.g., 5,10-methyleneTHF), or methanol (e.g., 5-methylTHF)
(Appling, 1991; Girgis et al., 1997; Schirch and Strong, 1989; Wagner, 1995).
Folate derivatives also contain a covalently bound polyglutamate peptide of
varying length. Serum folates contain a single glutamate residue, whereas
intracellular folates contain a polyglutamate peptide usually consisting of five
to eight glutamate residues that are polymerized through unusual g-linked\
peptide bonds (Moran, 1999; Shane, 1995). The polyglutamate pep-
tide increases the affinity of THF cofactors for folate-dependent enzymes
and-binding proteins, and prevents their efflux from the cell and intracellular
organelles (Schirch and Strong, 1989). THF polyglutamates are coenzymes
that donate or accept one-carbon units in a network of reactions known as
one-carbon metabolism that occurs in three specific and isolated cellular
compartments: the mitochondria, nucleus, and cytoplasm (Fig. 1.1; Porter
et al., 1985; Shane, 1989; Woeller et al., 2007a). The one-carbon forms of
THF can be interconverted enzymatically (Fig. 1.1), although each cofactor
form is specific to a particular biosynthetic pathway. The formyl group of
10-formylTHF is incorporated into the C2 and C8 of the purine ring
in the cytoplasm and is used to synthesize formylated methionyl-tRNA
in mitochondria (Fig. 1.1). The one-carbon moiety of 5,10-methyleneTHF
is required to convert uridylate to thymidylate, and the one carbon
carried by 5-methylTHF is required to remethylate homocysteine to
methionine. The cellular concentration of folate-binding proteins exceeds
that of folate derivatives, and therefore the concentration of free folate
in the cell is negligible (Schirch and Strong, 1989; Strong et al., 1990;
Suh et al., 2001). This implies that each folate-dependent biosynthetic
pathway competes for a limiting pool of folate cofactors (Scott et al., 1981;
Suh et al., 2001).
Epidemiological studies implicate impaired folate metabolism in several
pathologies and developmental anomalies including neural tube defects
(NTDs) (Scott, 2001; van der Put and Blom, 2000), cardiovascular disease
(Gerhard and Duell, 1999; Lindenbaumand Allen, 1995; Uelandet al., 2000),
2 Jennifer T. Fox and Patrick J. Stover
10-FormylTHF
5,10-MethenylTHF
5,10-MethyleneTHF
5-MethylTHF
Homocysteine
Thymidylate
Purines
Methionine
AdoMet
AdoHcy
Methylation reactions
Formate
glycine
Histidine
Purines
Serine
THF
THF
THF
5-formylTHF
5-MethylTHF
Sequestered
10-FormylTHF 5,10-MethenylTHF
5,10-MethyleneTHF
Serine
Glycine
CO
2
,NH
3
Glycine
Glycine
Dimethylglycine
Sarcosine
Sarcosine
fMet-tRNA Formate
Mitochondria Cytoplasm
1
2 3
4
5
6
7 8 9
11,12
13
14,15
16
16
17
17
19
20
16,22
21
DHF THF
5,10-MethyleneTHF
Glycine
Serine
dUMP
dTMP
Nucleus
23
16
19
THF
CO
2
10
THF
Figure 1.1 Compartmentation of folate-mediated one-carbon metabolism in the cytoplasm, mitochondria, and nucleus. One-carbon
metabolism in the cytoplasm is required for the de novo synthesis of purines and thymidylate and for the remethylation of homocysteine to
methionine. One-carbon metabolism in mitochondria generates one-carbon units for cytoplasmic one-carbon metabolism by generating
formate from serine, glycine, sarcosine, and dimethylglycine. One-carbon metabolism in the nucleus synthesizes dTMP from dUMP and
serine. 1, Mitochondrial serine hydroxymethyltransferase; 2, Aminomethyltransferase; 3, Sarcosine dehydrogenase; 4, Dimethylglycine
dehydrogense; 5, 5,10-Methylenetetrahydrofolate dehydrogenase (NAD-dependent); 6, 5,10-Methenyltetrahydrofolate cyclohydrolase;
and cancer (Ames, 2001; Blount et al., 1997; Choi and Mason, 2000; Kim,
1999; Pogribny et al., 1995). One-carbon metabolism can be impaired by
folate and other vitamin B deficiencies and/or common, penetrant genetic
mutations and polymorphisms (Bailey, 1995; McNulty, 1995; Scott, 1998;
van der Put and Blom, 2000). However, the biochemical mechanisms and
causal metabolic pathways responsible for the initiationand/or progressionof
folate-associated pathologies have yet to be established. In fact, there are still
major gaps in our fundamental understanding of one-carbon metabolismand
its regulation, including the potential for identifying putative missing
enzymes and their associated genes whose discovery may be necessary to
complete the assembly of the folate-dependent metabolic network. This
chapter focuses on our current understanding of mammalian folate-mediated
one-carbon metabolism, its cellular compartmentation, and knowledge gaps
that limit our understanding of folate metabolism and its regulation.
II. Introduction to Cytoplasmic One-Carbon
Metabolism
Folate-mediated one-carbon metabolism in the cytoplasm is a meta-
bolic network of interdependent biosynthetic pathways that are required for
the biosynthesis of purines and thymidylate, and the remethylation of
homocysteine to methionine (Fig. 1.1). Methionine can be adenosylated
to S-adenosyl methionine (AdoMet), a cofactor, and methyl group donor for
numerous methylation reactions including the methylation of neurotransmit-
ters and other small molecules, phospholipids, proteins including histones,
RNA, and cytosine bases within CpG islands in DNA. Many AdoMet-
dependent methylation reactions, including those involved in chromatin
methylation, serve regulatory functions by affecting gene transcription
(Miranda and Jones, 2007), protein localization (Winter-Vann et al., 2003),
and the catabolism of small molecules (Stead et al., 2004). The sources of
one-carbon moieties for cytoplasmic one-carbon metabolism include
formate, serine, histidine, and purines.
7, Methionyl-tRNA formyltransferase; 8, 10-Formyltetrahydrofolate synthetase;
9, 10-Formyltetrahydrofolate synthetase; 10, 10-Formyltetrahydofolate dehydroge-
nase; 11 and 12, Phosphoribosylglycinamide formyltransferase and Phosphoribo-
sylaminoimidazolecarboxamide formyltransferase; 13, 5,10-Methenyltetrahydrofolate
cyclohydrolase; 14 and 15, Glycine formiminotransferase/formimidoyltetrahydrofolate
cyclodeaminase andGlutamate formiminotransferase/formimidoyltetrahydrofolate cyclo-
deaminase; 16, Cytoplasmic serine hydroxymethyltransferase; 17, Methenyltetrahydrofo-
late synthetase; 18, 5,10-Methylenetetrahydrofolate dehydrogenase (NADP-dependent);
19, Thymidylate synthase; 20, Methylenetetrahydrofolate reductase; 21, Methionine
synthase; 22, Glycine N-methyltransferase; 23, Dihydrofolate reductase.
4 Jennifer T. Fox and Patrick J. Stover
Proteins involved in folate metabolism can be classified into four func-
tional categories, although many folate-dependent enzymes exhibit two or
more of these activities: (1) One-carbon generating enzymes, (2) THF one-
carbon interconverting enzymes, (3) THF-dependent biosynthetic enzymes,
and (4) noncatalytic THF-binding proteins (Table 1.1). This section details
the mechanisms, regulation, and physiological functions of the enzymes that
are involved in cytoplasmic one-carbon metabolism, as well as common
genetic variants that affect enzyme function and the one-carbon network.
A. Enzymes that generate one-carbon units
1. Cytoplasmic serine hydroxymethyltransferase
a. Reaction Serine hydroxymethyltransferase (SHMT) catalyzes the
reversible and PLP-dependent interconversion of serine and glycine. Mam-
mals express cytoplasmic (SHMT1) and mitochondrial (SHMT2) SHMT
isozymes; the human isozymes share 63% amino acid sequence identity
and are encoded on separate genes (Garrow et al., 1993). When catalyzing
serine cleavage, SHMT1 transfers the C3 of serine to THF generating
glycine and 5,10-methyleneTHF, the cofactor required for thymidylate
biosynthesis. The one-carbon moiety of 5,10-methyleneTHF can also
support homocysteine remethylation when converted to 5-methylTHF
by methylenetetrahydrofolate reductase (MTHFR) (Fig. 1.1). SHMT1-
derived one-carbons are not believed to make significant contributions to
purine biosynthesis because the reductive environment (NADPH/
NADP
H
2
O
tetrahydrofolate CO
2
NADPH H
Cytoplasmic serine
hydroxymethyltransferase
2.1.2.1 Cytoplasm and
nucleus
5,10-Methylenetetrahydrofolate glycine H
2
O
tetrahydrofolate L-serine
Glycine formiminotransferase 2.1.2.4 Cytoplasm N-formiminoglycine tetrahydrofolate )glycine
5-formiminotetrahydrofolate
Glutamate
formiminotransferase
2.1.2.5 Cytoplasm N-formiminoglutamate tetrahydrofolate )glutamate
5-formiminotetrahydrofolate
Mitochondrial serine
hydroxymethyltransferase
2.1.2.1 Mirochondrian 5,10-Methylenetetrahydrofolate glycine H
2
O
tetrahydrofolate L-serine
Sarcosine dehydrogenase 1.5.99.1 Mitochondrian Sarcosine acceptor H
2
O glycine formaldehyde
reduced acceptor
Dimethylglycine
dehydrogense
1.5.99.2 Mitochondrian N,N-dimethylglycine acceptor H
2
O sarcosine
formaldehyde reduced acceptor
Aminomethyltransferase 2.1.2.10 Mitochondrian [Protein]-S8-aminomethyldihydrolipoyllysine
tetrahydrofolate [protein]-dihydrolipoyllysine
5,10-methylenetetrahydrofolate NH
3
One-carbon interconverting enzymes
5,10-Methylenetetrahydrofolate
dehydrogenase (NADP-
dependent)
1.5.1.5 Cytoplasm 5,10-Methylenetetrahydrofolate NADP
5,10-
methenyltetrahydrofolate NADPH H
6
5,10-Methylenetetrahydrofolate
dehydrogenase (NAD-
dependent)
1.5.1.15 Mitochondrian 5,10-Methylenetetrahydrofolate NAD
5,10-
methenyltetrahydrofolate NADH H
5,10-Methenyltetrahydrofolate
cyclohydrolase
3.5.4.9 Cytoplasm 5,10-Methenyltetrahydrofolate H
2
O
10-formyltetrahydrofolate
Methenyltetrahydrofolate
synthetase
6.3.3.2 Cytoplasm ATP 5-formyltetrahydrofolate ADP phosphate 5,10-
methenyltetrahydrofolate
Methylenetetrahydrofolate
reductase
1.5.1.20 Cytoplasm 5-Methyltetrahydrofolate NAD(P)
5,10-
methylenetetrahydrofolate NAD(P)H H
7,8-dihydrofolate
NADPH H
Formimidoyltetrahydrofolate
cyclodeaminase
4.3.1.4 Cytoplasm 5-Formimidoyltetrahydrofolate 5,10-
methenyltetrahydrofolate NH
3
Biosynthetic enzymes
Phosphoribosylglycinamide
formyltransferase
2.1.2.2 Cytoplasm 10-Formyltetrahydrofolate N1-(5-phospho-D-ribosyl)
glycinamide tetrahydrofolate N2-formyl-N1-
(5-phospho-D-ribosyl)glycinamide
Phosphoribosylaminoi
midazolecarboxamide
formyltransferase
2.1.2.3 Cytoplasm 10-Formyltetrahydrofolate 5-amino-1-(5-phospho-
D-ribosyl)imidazole-4-carboxamide tetrahydrofolate
5-formamido-1-(5-phospho-D-ribosyl)imidazole-
4-carboxamide
Methionine synthase 2.1.1.13 Cytoplasm 5-Methyltetrahydrofolate L-homocysteine
tetrahydrofolate L-methionine
Thymidylate synthase 2.1.1.45 Cytoplasm and
nucleus
5,10-Methylenetetrahydrofolate dUMP dihydrofolate
dTMP
Methionyl-tRNA
formyltransferase
2.1.2.9 Mitochondrian 10-Formyltetrahydrofolate L-methionyl-tRNAfMet H
2
O
tetrahydrofolate N-formylmethionyl-tRNAfMet
Other folate-binding proteins
Glycine N-methyltransferase 2.1.1.20 Cytoplasm S-adenosyl-L-methionine glycine S-adenosyl-L-
homocysteine sarcosine
7
condenses with THF to form 5,10-methyleneTHF through an N5 imi-
nium cation intermediate. However, this mechanism is not consistent
with the structure of the Bacillus stearothermophilus SHMT (bsSHMT)
serine complex (Trivedi et al., 2002) and metabolic labeling experi-
ments (Tatum et al., 1977), both of which indicate that the C3-hydroxyl
group of serine is oriented in the synperiplanar configuration rather than
the antiperiplanar configuration required for a retroaldol mechanism.
In addition, the catalytic base has never been identified. A second puta-
tive direct displacement mechanism was revealed from the structures of
bsSHMT complexed with serine and with 5-formylTHF and glycine
(Trivedi et al., 2002). This proposed mechanism proceeds with the N5
of THF displacing the C3 hydroxyl of serine to form a covalent interme-
diate. However, Szebenyi et al. provide evidence that the reverse reaction
could not proceed by direct displacement and that the position of N5
of THF is unfavorable for nucleophilic attack. Rather, they propose a
third mechanism whereby the N5 of THF attacks the C3-hydroxy
group of serine to form N5-hydroxymethyleneTHF, glycine, and possibly
a transient formaldehyde intermediate (Schirch and Szebenyi, 2005;
Szebenyi et al., 2004).
c. Regulation In addition to their primary catalytic function, both SHMT
isozymes catalyze the irreversible conversion of 5,10-methenylTHF to
5-formylTHF (Fig. 1.1). 5-FormylTHF is not a cofactor for folate-
dependent one-carbon transfer reactions, but rather is an inhibitor of several
folate-dependent reactions including SHMT (Stover and Schirch, 1990,
1993). The SHMT isozymes may also play roles as folate-binding proteins
(Stover and Schirch, 1991). Both 5-formylTHF and 5-methylTHF poly-
glutamates are tight-binding SHMT inhibitors.
Unlike many enzymes involved in cytoplasmic folate-mediated one-
carbon metabolism, SHMT1 is not ubiquitously expressed in tissues, but it is
abundant in the liver, kidney, and colon and is also found in the brain
(Girgis et al., 1998). Its expression and/or activity are regulated by several
nutrients and metabolic factors including pyridoxal phosphate (vitamin B
6
),
retinoic acid, zinc, and ferritin. Vitamin B
6
deficiency was shown to
decrease SHMT1 activity in rat liver (Scheer et al., 2005) and protein levels
in cultured cells (Perry et al., 2007). Retinoic acid, which inhibits prolifera-
tion and induces differentiation during vertebrate development, greatly
reduces SHMT1 mRNA levels (Nakshatri et al., 1996). In contrast, zinc
induces SHMT1 transcription by acting through a metal regulatory element
present within the promoter (Perry et al., 2005). The heavy chain subunit
of the iron-storage protein ferritin was also shown to increase SHMT1
protein levels (Oppenheim et al., 2001) by stimulating the cap-independent
translation of the transcript (Woeller et al., 2007b).
8 Jennifer T. Fox and Patrick J. Stover
d. Physiological function/gene variants Although the SHMT1 and
SHMT2 isozymes exhibit similar catalytic and physical properties, they
have distinct physiological functions. Loss of the mitochondrial SHMT2
isozyme creates a glycine auxotrophy in Chinese hamster ovary cells
(Chasin et al., 1974), indicating that SHMT1 is not a primary source of
glycine and that SHMT1 cannot substitute for SHMT2 function. Stable
isotope tracer studies using cultured cells indicate that SHMT1-derived
5,10-methyleneTHF is preferentially directed to thymidylate biosynthesis
relative to homocysteine remethylation (Herbig et al., 2002; Oppenheim
et al., 2001). This preferential partitioning of SHMT1-derived one-carbons
to thymidylate synthesis may be achieved through the cell cycle-dependent
partitioning of the thymidylate synthesis pathway in the nucleus (Anderson
et al., 2007) (Fig. 1.1; see Section IV). The SHMT1 protein has also been
demonstrated to be a 5-methylTHF tight-binding protein in cultured cells;
increased expression of SHMT1 increased cellular levels of 5-methylTHF at
the expense of other one-carbon forms of folate while depleting AdoMet
levels (Herbig et al., 2002). This latter observation is consistent with
SHMT1 serving as a 5-methylTHF-binding protein in the cytoplasm and
thereby limiting the availability of 5-methylTHF for homocysteine
remethylation (Fig. 1.1; Herbig et al., 2002).
A common SHMT1 single nucleotide polymorphism (SNP), C1420T,
has been shown to be protective against adult acute lymphocytic leukemia
(ALL) (Skibola et al., 2002) and malignant lymphoma (Hishida et al., 2003).
This SNP results in an amino acid substitution of leucine to phenylalanine at
position 474 of the protein (L474F) and prevents SHMT1 SUMOylation
(Woeller et al., 2007a). The SHMT1 C1420T gene variant has also been
shown to be associated with elevated plasma and red cell folate levels (Heil
et al., 2001), and some studies report that it protects against NTDs (Relton
et al., 2004a,b). When present in combination with the MTHFR C677T
polymorphism (see Section II.B.4), SHMT1 C1420T is a risk factor for
cardiovascular disease (Lim et al., 2005).
2. 10-FormylTHF synthetase
a. Reaction 10-FormylTHF synthetase (FTHFS) is a formate-activating
enzyme found in a wide variety of organisms including bacteria, plants, insects,
nematodoes, yeast, and mammals. In eukaryotes, FTHFS activity is found on
the C-terminal domain of the trifunctional enzyme, C
1
-THF synthase, which
also contains 5,10-methenylTHF cyclohydrolase (MTHFC) and 5,10-methy-
leneTHF dehydrogenase (MTHFD) activities (see Section II:B.1; Howard
et al., 2003). C
1
-THF synthase is encoded by the Mthfd1 gene. FTHFS
catalyzes the ATP-dependent conversion of THF and formate to
10-formylTHF, ADP, and inorganic phosphate. The reaction is reversible
(Appling, 1991). The enzyme requires monovalent cations (NH
4
, K
, or
Rb
-depen-
dent aldehyde-dehydrogenase reaction, and the N-terminal domain cata-
lyzes the hydrolysis of 10-formylTHF to THF and formate (hydrolase
reaction) (Cook et al., 1991; Donato et al., 2007). Although the two
domains can function independently, the two active sites work in concert
through a 4
0
-phosphopantetheine swinging arm that is bound through a
phosphoester bond to Ser354; the swinging arm transfers formate between
the two active sites (Donato et al., 2007).
b. Mechanism The FDH reaction initiates with a hydrolase reaction
where a water molecule, activated by aspartate 142, acts as a nucleophile
by attacking the formyl carbon atom of 10-formylTHF to produce a
hydrated aldehyde intermediate (Donato et al., 2007; Tsybovsky et al.,
2007). In the absence of NADP
and C. M. Ulrich
Contents
I. Introduction 46
II. Structure and Function of the Cycles 49
III. Why Mathematical Modeling? 51
A. Previous modeling efforts 52
B. Why modeling? 54
C. Difficult issues in modeling 55
D. Advantages of mathematical models 57
E. Kinetics, parameter values, and model structure 59
IV. Model Development 61
V. Blood Versus Intracellular Metabolite Concentrations 66
VI. Modeling GeneGene and GeneEnvironment Interactions 67
VII. Modeling and Simulation have Revealed Novel Homeostatic
Mechanisms 70
VIII. Steady States and Fluctuations 75
IX. Conclusions 77
Acknowledgments 78
References 78
Abstract
Folate-mediated one-carbon metabolism is an unusually complex metabolic
network, consisting of several interlocking cycles, and compartmentation
between cytosol and mitochondria. The cycles have diverse functions, the
primary being thymidylate synthesis (the rate limiting step in DNA synthesis),
the initial steps in purine synthesis, glutathione synthesis, and a host of methyl
transfer reactions that include DNA and histone methylation. Regulation within
the network is accomplished by numerous allosteric interactions in which
Vitamins and Hormones, Volume 79
#
2008 Elsevier Inc.
ISSN 0083-6729, DOI: 10.1016/S0083-6729(08)00402-0 All rights reserved.
* Department of Biology, Duke University, Durham, North Carolina 27705
{
Department of Mathematics, Duke University, Durham, North Carolina 27705
{
Cancer Prevention Program, Fred Hutchinson Cancer Research Center, Seattle, Washington 98109
45
metabolites in one part of the network affect the activity of enzymes elsewhere
in the network. Although a large body of experimental work has elucidated the
details of the mechanisms in every part of the network, the multitude of
complex and non-linear interactions within the network makes it difficult to
deduce how the network as a whole operates. Understanding the operation of
this network is further complicated by the fact that human populations maintain
functional polymorphisms for several enzymes in the network, and that the
network is subject to continual short and long-term fluctuations in its inputs as
well as in demands on its various outputs. Understanding how such a complex
system operates is possible only by means of mathematical models that take
account of all the reactions and interactions. Simulations with such models can
be used as an adjunct to laboratory experimentation to test ideas and alterna-
tive hypotheses and interpretations quickly and inexpensively. A number of
mathematical models have been developed over the years, largely motivated
by the need to understand the complex mechanisms by which anticancer drugs
like methotrexate inhibit nucleotide synthesis and thus limit the ability of cells
to divide. More recently, mathematical models have been used to investigate
the regulatory and homeostatic mechanisms that allow the system to accom-
modate large fluctuations in one part of the network without affecting critical
functions elsewhere in the network. 2008 Elsevier Inc.
I. Introduction
The origin of our mathematical modeling work stems froman interest in
understanding howgenes and the environment interact in the biochemistry of
cells. This ledus tostudy folate andmethionine metabolismbecause this part of
cell metabolismis linked to a diversity of humandiseases that have bothgenetic
and environmental contributing factors. Folate and other B vitamins play
critical roles in the biochemical reactions of one-carbon metabolism that are
related to amino acid metabolism, nucleotide synthesis, and numerous
methyl-transferase reactions, including DNA and protein methylation.
Defects in folate-mediated one-carbon metabolism (FOCM; Table 2.1
lists the acronyms and abbreviation used in this chapter), either due to
mutations in the genes that code for enzymes in the pathway or to deficien-
cies in vitamin cofactors, are associated with megaloblastic anemia, spina
bifida and other neural tube defects, cardiovascular disease, increased sensi-
tivity to oxidative stress, and a variety of neuropsychiatric disorders. FOCM
is also involved in the etiology of colorectal and other types of cancer, and
chemotherapeutic agents, such as methotrexate and 5-fluorouracil, target
FOCM and play a central role in cancer treatment. FOCM is highly com-
plex. It consists of a set of interlocked biochemical cycles (Fig. 2.1) whose
enzymes are subject to complex allosteric regulations. The function of this
complex network is further complicated by the fact that there are genetic
polymorphisms for many of the enzymes in the network, and the functions
46 H. F. Nijhout et al.
of the network are sensitive to the input of various amino acids (glycine,
serine, methionine, and cysteine), B vitamins (folic acid, B
6
and B
12
), and is
affected by environmental factors such as alcohol intake in intricate ways that
alter the normal operation of the network and the risk of disease.
Considerable research over the past 40 years has identified most if not all of
the important details of FOCM. However, a limiting factor of these critical
studies is that they have primarily focused on single reactions and on small
portions of the pathway, and thus provide no means for understanding the
overall functioning of the system. The multiple cycles andpathways of FOCM
together are part of a complex nonlinear system, which is difficult to capture
using purely experimental methods. Mathematical modeling is an approach
that has been particularly useful in the study of complex biological systems
(Edelstein-Keshet, 1988; Murray, 1989). Below we will review how mathe-
matical modeling has been able to confirmkey hypotheses about the operation
of various portions of FOCM, and how modeling has provided novel insights
into the properties and consequences of various regulatory mechanisms that
stabilize portions of the network against environmental perturbations.
Table 2.1 Abbreviations and acronyms used in the text and gures
Acronym Name
10f-THF 10-Formyltetrahydrofolate
10f0DHF 10-Formyldihydrofolate
5fTHF 5-Formyltetrahydrofolate (leucovorin)
5mTHF 5-Methyltetrahydrofolate
AICAR P-ribosyl-5-amino-4-imidazole carboxamide
AICART Aminoimidazolecarboxamide ribonucleotide transferase
BET, Bet Betaine
BHMT Betaine-homocysteine methyltransferase
CBS Cystathionine b-synthase
CH:NHTHF 5-Formiminotetrahydrofolate
CHTHF 510-Methenyltetrahydrofolate
CH2-THF 510-Methylenetetrahydrofolate
CTGL g-Cystathionase
Cys Cysteine
Cyst Cystathionine
DHF Dihydrofolate
DHFR Dihydrofolate reductase
DHFS Dihydrofolate synthase
DHPR Dihydropteridine reductase
DMG Dimethylglycine
DMGD Dimethylglycine dehydrogenase
DNMT DNA-methyltransferase
(continued)
Mathematical Models of Folate-Mediated One-Carbon Metabolism 47
Table 2.1 (continued)
Acronym Name
dTMP Deoxythymidine monophophate
dUMP Deoxyuridine monophophate
FOCM Folate-mediated one-carbon metabolism
FR-RFC Folate receptor reduced folate carrier
FTD 10-Formyltetrahydrofolate dehydrogenase
FTD 10-Formyltetrahydrofolate dehydrogenase
FTS 10-Formyltetrahydrofolate synthase
GAR Glycinamide ribonucleotide
GCS g-Glutamylcysteine synthetase
GDC Glycine decarboxylase (glycine cleavage system)
Glut Glutamate
Glut-Cys Glutamyl-cysteine
Gly Glycine
GNMT Glycine N-methyltransferase
GPX Glutathione peroxidase
GR Glutathione reductase
GS Glutathione synthetase
GSH Reduced glutathione
GSSG Oxidized glutathione disulde
H
2
CO Formaldehyde
H
2
O
2
Hydrogen peroxide
HCOOH Formate
Hcy Homocysteine
MAT-I Methionine adenosyl transferase I
MAT-II Methionine adenosyl transferase II
MAT-III Methionine adenosyl transferase III
Met Methionine
MS Methionine synthase
MTCH 5,10-Methenyltetrahydrofolate cyclohydrolase
MTD 5,10-Methylenetetrahydrofolate dehydrogenase
MTHFR 5,10-Methylenetetrahydrofolate reductase
MTS 5,10-Methenyltetrahydrofolate synthetase
NADPH Nicotinamide adenine dinucleotide phosphate
NE non-enzymatic conversion
PGT Phosphoribosyl glycinamidetransformylase
SAH S-adenosylhomocysteine
SAHH S-adenosylhomocysteine hydrolase
SAM S-adenosylmethionine
Sarc Sarcosine
SDH Sarcosine dehydrogenase
Ser Serine
SHMT Serinehydroxymethyltransferase
TS Thymidylate synthase
THF Tetrahydrofolate
48 H. F. Nijhout et al.
II. Structure and Function of the Cycles
FOCM consists of three functional modules: the folate cycle, the
methionine cycle, and the glutathione synthesis pathway. In this chapter,
we do not consider the role of the polyamine synthesis pathway which has
recently been modeled by Rodriguez-Caso et al. (2006).
The function of FOCM is to pick up carbons from amino acids, primarily
serine, but also glycine and methionine, and deliver them (as methyl groups)
for the synthesis of purines and pyrimidines and for a variety of methylation
reactions (e.g., DNA, tRNA, and histones) (Clarke and Banfield, 2001; Cook,
2001; Shane, 1995; Wagner, 1995). Serine enters the reactions as a substrate
for SHMT which transfers one carbon to THF, yielding glycine and CH
2
-
THF. The glycine can enter the mitochondria where it is processed by the
glycine cleavage system, transferring one of its carbons to CH
2
-THF in the
THF
CH2-THF
DHF
10f-THF
dUMP
Pyrimidine
synthesis
dTMP
CH=THF
MTCH
FTS
NE
Purine
synthesis
DHFR GAR
FTD
ATP
Betaine
H
2
O
H
2
C=O
5mTHF
DNA
DNA-CH
3
Methylation
reactions
SAM
SAH Hcy
Met
Adenosine
cystathionine
BHMT
GNMT
DNMT
MAT-I
MAT-III
SAHH
CBS
MTHFR
MTD
PGT
AICART
MS
TS
FTD
H
2
C=O
Mitochondria
CO
2
FTS
NE
DMGD
SDH
GDC
SHMT
10f-THF
CH=THF
CH2-THF
THF
MTD
MTCH
Gly
Gly Gly
Gly
Ser
Ser Ser
Sarc
DMG
DMG
Sarc
SHMT
Cytosol
Gluconeogenesis
HCOOH HCOOH
vHCOOH
vSer
vGly
Serin Glyin
Metin
Serine Glycine
Methionine
SerOut
AICAR
Glutamate
Cystathionine
Cys
Glut-Cys
CTGL
GCS
GS
GSH
Gly
Cysteine
GSSG
GPX
GR
Glutathione(GSH) GSSG
Cysin
Oxidative stress
H
2
O
2
Blood
Blood
FR-RFC
Folate
One-Carbon Metabolism
Figure 2.1 Diagram of mammalian hepatic folate-mediated one-carbon metabolism.
This pathway includes the mitochondrial compartmentation, reduced glutathione
synthesis, and transport of some metabolites from the blood. Metabolites that are
variables in the model are enclosed in boxes, and enzymes are in ellipses. Full names
of the acronyms of enzymes and metabolites are given in Table 2.1.
Mathematical Models of Folate-Mediated One-Carbon Metabolism 49
mitochondrial folate cycle. The mitochondrial cycle then releases its carbons
to the cytosol as formate (HCOOH in Fig. 2.1). Glycine is also used in the
synthesis of sarcosine by GNMT. Sarcosine, in turn, also enters the mitochon-
dria and eventually yields all but one of its carbons to the mitochondrial folate
cycle. Serine is also used by the CBS reaction which complexes it with
homocysteine to yield cystathionine, which, in turn, is used for the synthesis
of cysteine and glutathione. The one-carbon units held by CH
2
-THF have
three fates: they can be passed to 5mTHF by MTHFRand subsequently to the
methionine cycle where they are used in a great diversity of methylation
reactions; they can also be passed to 10f-THF and subsequently used for
purine synthesis; finally they can be passed to TS and used to synthesize
dTMP fromdUMP. Thus, FOCMplays a critical role in nucleotide synthesis,
and the TS reaction is the rate-limiting step in DNA synthesis (Fukushima
et al., 2003). The cytosolic and mitochondrial SHMT reactions are reversible.
In the forward direction they use serine, and in the backward direction they
use glycine and one-carbon units from the mitochondrial folate cycle to
synthesize serine, which canserve as the basis for gluconeogenesis. Methionine
enters the methionine cycle and is adenosylated by MAT-I and MAT-III
(in the liver; MAT-II is the adenosyl transferase used in other tissues).
S-adenosyl methionine (SAM) serves as the general methyl donor for the
majority of methylation reactions in the cell. About half of the mass of
methionine that enters the methionine cycle leaves via the transulfuration
pathway to cystathionine and cysteine (Finkelstein, 1990; Finkelstein and
Martin, 1986), and the other half is remethylated to methionine by MS and
BHMT, using methyl groups from 5mTHF and betaine, respectively.
If the reactions illustrated in Fig. 2.1 were the only pertinent ones, this
would be a case of complicated but standard biochemistry. However, many
of the metabolites in this system are allosteric activators or inhibitors of
enzymes at some distance in the network. For example, SAM inhibits
BHMT and MTHFR and activates CBS; DHF inhibits MTD, MTCH,
and MTHFR; 5mTHF inhibits GNMT and SHMT (Finkelstein, 2003;
Finkelstein and Martin, 1984; Finkelstein et al., 1972; Jencks and Matthews,
1987; Kluijtmans et al., 1996; Ou et al., 2007; Yamada et al., 2001; Yeo and
Wagner, 1992). Many reactions also depend strongly on the cellular status
for folate and vitamins B
6
and B
12
. In addition, the velocities of many
reactions depend on the concentrations of the substrates that are controlled
by dietary inputs of glycine, serine, glutamate, cysteine and methionine.
These inputs naturally undergo enormous fluctuations so the system is often
far away from steady state (Nijhout et al., 2007b).
It is a significant challenge to understand the biological reasons for the
complicated interlocking cycles, the compartmentalization to the mitochon-
dria, and the multiple reactions by which one substrate can be transformed
into another. Since many parts of the network of FOCM share enzymes and
metabolites, there must be mechanisms that ensure that large variation in a
50 H. F. Nijhout et al.
particular region of the network does not compromise the function in other
regions. Furthermore, the many critical reactions in the network must be
buffered against large and irregular hourly and daily fluctuations in inputs of
amino acids. The overall network is too complex to understand these
regulatory functions by inspection of the reaction diagram alone, or to
deduce the integration of its various functions with any degree of certainty.
The system is well-enough understood that it is possible to develop a
mathematical description of the kinetics of the various individual reactions,
and couple these together into a single mathematical simulation model that
can be used to explore questions about structure and function.
Over the past 5 years, we have developed mathematical models for the
pathways shown in Fig. 2.1, which represents mammalian hepatic one-
carbon metabolism. Hepatic FOCM is what we might call the complete
pathway, in that all known enzymes and reactions operate in the liver. This
is not true for most other tissues. While most FOCM enzymes are also
expressed in the kidney, most other tissues in the body express only a subset
of the enzymes and thus operate on what we might called reduced
FOCM. FOCM is an ancient pathway and we have recently developed
a model for the structure and kinetics of folate metabolism in bacteria
(Leduc et al., 2007 and Fig. 2.6).
III. Why Mathematical Modeling?
We view a mathematical model as an experimental tool, much like
electrophoresis or PCR or gene knockout. Like all experimental tools,
models have their own particular strengths and limitations and these should
be understood if the tool is used to address a particular problem. A model is a
mathematical description of a specific system. One of the particular
strengths of a model is that it is completely explicit about what is in the
system and what is not. In addition, a model is explicit about all
the assumptions that are made about the properties of the components of
the system and about their interactions.
The mathematical models we are dealing with here are not theoretical
models in the sense that they attempt to discover necessary and sufficient
conditions for the behavior of a particular system, or attempt to estimate
parameter values for the system. Rather, they are strict quantitative descrip-
tions of properties that have been determined experimentally by investiga-
tors. Ideally, a model represents the state of our understanding of the
properties and interactions among the component parts of a system, and
allows one to examine the behavior of the ensemble, and the consequences
for the system as a whole of various assumptions one makes about how the
components behave and interact. The model is tested against as much
Mathematical Models of Folate-Mediated One-Carbon Metabolism 51
experimental data as possible. Ideally, the model reproduces the results of a
broad diversity of experiments both qualitatively and quantitatively.
If it does, then it can be used as an experimental tool to ask questions and
do experiments that would be difficult, or expensive, or unethical to do
in a real living system. In particular, a model provides a means to rapidly and
inexpensively test the effects of specific perturbations and of alternative
experimental strategies before committing time and resources to potentially
expensive, and possibly inconclusive laboratory experiments. In its most
useful guise, simulations with a model should interact with laboratory
experimentation in a mutually illuminating exploration of FOCM.
A modeling approach is useful when the system one wishes to study is
large and complex, with nonlinear interactions. Nonlinear systems produce
context-dependent and nonintuitive responses to perturbations, and a sim-
ple examination of the connectivity diagram is seldom able to reveal
anything useful about the dynamics of the system, nor its response to
perturbation. FOCM is such a large, complex, and nonlinear system, con-
sisting of several interlocking cycles with multiple inputs and outputs
(Fig. 2.1). In addition, many of the enzymes in this system are subject to
complex allosteric regulation by metabolites that are many steps removed in
the network. These long-range regulatory interactions provide important
homeostatic functions (see Section VII), which can only be evaluated by
simulation studies.
A. Previous modeling efforts
A number of investigators have developed mathematical modes for various
parts of FOCM (Harvey and Dev, 1975; Jackson and Harrap, 1973;
Morrison and Allegra, 1989; Seither et al., 1989; Vorontzov et al., 1980;
Werkheiser et al., 1973) and the methionine cycle (Martinov et al., 2000;
Prudova et al., 2005). Perhaps the best known among these are the extensive
studies of R. C. Jackson and his associates ( Jackson, 1980, 1984, 1986,
1993, 1995; Jackson and Harrap, 1973, 1979).
Almost without exception the models of folate metabolism have been
aimed at understanding the mechanism of action and the kinetics of anti-
cancer drugs, particularly methotrexate and 5-fluorouracil. In most cases the
models focused on the portions of the system that were most relevant for
their investigations. Jackson (1980, 1986; Jackson and Harrap, 1979) devel-
oped what is probably the most extensive model for folate metabolism
(Fig. 2.2A), consisting of more than 60 reactions that also included the
kinetics of membrane transport of folate and methotrexate, and more
detailed reactions for the synthesis of purines, pyrimidines, RNA, and
DNA. This model made specific predictions about the rates of DNA
synthesis and the amount of time required to replicate all the DNA in a
cell, and was thus able to estimate the maximal rate of cell division under
52 H. F. Nijhout et al.
various methotrexate treatment regimes. The simulated results closely
matched experimentally observed data on the inhibition of cell division
by various methotrexate treatments. The Morrison and Allegra model
(1989), shown in Fig. 2.2B, dealt specifically with the kinetics of folate
metabolismin the MCF-7 breast cancer cell line, and included the effects of
methotrexate polyglutamation and the consequent improved cellular reten-
tion of methotrexate. In spite of the fact that these models typically dealt
only with subsections of one-carbon metabolism, and did not include any of
the allosteric interactions that regulate and stabilize metabolite and fluxes,
Glycine
5mTHF
THF
5,10-CH2-THF
S
H
M
T
Serine
DHF
A
B
10f-THF
TS
MS
DHFR
dUMP
dTMP
Pyrimidine
synthesis
5,10-CH=THF
MTHFR
MTD
MTCH
FTS
NE
PGT
AICART
Purine
synthesis
AICAR
NADPH
NADPH
NADPH NADP
+
NADP
+
NADP
+
GAR
HCOOH
FTD
Methionine
Hcy
H
2
C=O
CH:NHTHF
NADPH
NADP
+
NADPH
NADPH
NADP
+
NADP
+
Glycine
5mTHF
THF
5,10-CH2-THF
S
H
M
T
Serine
10-fDHF
10f-THF
TS
MS
DHFR
dUMP
dTMP
Pyrimidine
synthesis
MTHFR
MTD
FTS
N
E
PGT
AICART
Purine
synthesis
AICAR
GAR
HCOOH
Methionine
Hcy
H
2
C=O
DHF
AICART
FDS
Figure 2.2 Diagrams of two early models of folate metabolism. A, the model of
Jackson (1980). This model also includes synthesis of nucleotides, RNA and DNA, as
well as the transport of folates and methotrexate into the cell, not shown in this
diagram. B, the model of Morrison and Allegra (1989). Full names of the acronyms
of enzymes and metabolites for this and other figures, and the text, are given in
Table 2.1. Redrawn from Jackson (1980) and Morrison and Allegra (1989).
Mathematical Models of Folate-Mediated One-Carbon Metabolism 53
they were generally able to simulate the correct pool sizes of several of the
metabolites, and the time course of inhibition of nucleotide synthesis rates
by treatment with antifolate cancer drugs.
B. Why modeling?
The traditional negative view of mathematical modeling is the following.
If the biology and biochemistry are well understood, then there is no reason
for models. On the other hand, if the biology or biochemistry is not well
understood then there is not enough information to make an accurate
model. Therefore, in either case, mathematical modeling is useless. And,
of course, this negative view is reinforced by poor modeling or modelers
who do not want deal with the full complexity of biological systems. In fact,
many biological systems are partly understood in the sense that there is
good information about many of the components of the systems but
incomplete information about how the components work together to
give rise to functional system properties. This is exactly the case with
FOCM where a great deal of information is available on individual reactions
but there is not much understanding of how the whole system works
together. It is in this intermediate partly understood situation that a
mathematical model can be a valuable, indeed necessary, investigative
tool. To illustrate this point, it is helpful to face how difficult it really is to
understand FOCM.
It is possible to understand a moderately sized traditional biochemical
reaction diagram by walking the diagram. If substrate A goes up then, since
A makes B, we expect B to go up and so forth. However, the existence of
allosteric interactions by which substrates in one part of the network activate
or inhibit enzymes in other parts makes this type of simple reasoning
impossible or at best inconclusive. For example, SAM activates the enzyme
CBS and inhibits the enzyme MTHFR. So, if we moderately increase the
methionine input to the system, will the homocysteine concentration go up
or down? Well, we would expect more mass in all the methionine cycle
metabolites, so [Hcy] should go up. On the other hand, when methionine
input goes up SAM rises appreciably and this will activate CBS, which will
draw down [Hcy]. But since SAM is up, it will inhibit MTHFR, which
will lower [5mTHF]. Since [5mTHF] is lower, the MS reaction will run
more slowly and thus [Hcy] is not used as rapidly and thus should go up. So
what will happen to [Hcy]? It is clear that no amount of verbal reasoning is
going to answer this question, especially since the reactions and the allosteric
interactions are nonlinear. One has to make calculations (and experiments)
about the relative strengths of the competing influences on [Hcy]. Two
other issues make the question even more daunting. First, the answer
may depend on the overall context of the rest of FOCM (see below).
Second, the allosteric activation of CBS and inhibition of MTHFR may
54 H. F. Nijhout et al.
have evolved to stabilize [Hcy] concentration in the face of moderate
changes in methionine input, in which case the answer to the question of
whether [Hcy] goes up or down is neither: it doesnt change much.
It is now well understood that gene expression is a stochastic process that
leads to phenotypic protein differences even among identical cells
(Elowitz et al., 2002; Sigal et al., 2006). Not only do the protein levels
vary by as much as 1530% from the mean from cell to cell, but also the
levels vary over time even in individual cells. This variation has conse-
quences for the understanding of FOCM. First, it does not make sense to
ask for the exact value of a given parameter (a V
max
, a K
m
, or a K
i
). Those
values will vary substantially from cell to cell and from time to time in any
given cell. Second, specific questions like Does [Hcy] go up or down?
may have answers that depend on the context of all the other enzymes in the
system. Third, some of the most important properties of FOCM (or indeed
of all of cell metabolism) are regulations, not obvious from the standard
biochemical reaction diagram, that allow the system to function despite
these large variations. The allosteric interactions mentioned above are
examples of such regulations. Thus, FOCM should not be thought of as a
single fixed system but a whole family of systems with large variations in
important parameters. It is difficult to see how one could understand
function in the face of such variation without mathematical modeling.
In Nijhout et al. (2007b), we show how many of the concentrations and
reaction velocities of hepatic FOCM react to the daily inputs of amino acids
due to meals. Some concentrations and velocities fluctuate wildly while
others are protected by regulatory mechanisms. More recent calculations
with the full model depicted in Fig. 2.1 (see Fig. 2.10) show the same
behavior. This is the reason, of course, that many experimental and clinical
measurements are done in the fasting state. Because of the difficulty of
making many simultaneous measurements of concentrations and velocities
as functions of time in living cells it is difficult to see how such dynamic
fluctuating behavior could be investigated experimentally. Thus, mathe-
matical modeling has a central role in elucidating the regulatory mechanisms
that allow cells to adapt to such dramatic changes in inputs.
C. Difficult issues in modeling
Suppose that one wants to investigate a particular phenomenon in FOCM
seen experimentally or clinically, for example, the behavior of homocyste-
ine under methionine loading or the stability of the glutathione pool in the
face of daily meals. Which variables and interactions should be included?
If the model is too small, one may have excluded (or rather held constant)
just those variables and interactions that are crucial to understanding the
phenomenon. If the model is too large it may be too unwieldy to experi-
ment with and the noise from all the approximations that one makes may
Mathematical Models of Folate-Mediated One-Carbon Metabolism 55
obscure the phenomenon that one wants to study. So, how large should a
model be? This is always a difficult question (though usually not admitted by
modelers), and every modeling attempt answers it explicitly by what is
included and what is excluded. Since our goal is not only to reproduce
experimental or clinical results but also to use the model to understand how
and why they arise, our philosophy is to start with smaller models and
expand to larger models when the smaller ones are well understood and the
expansion to more variables is necessary. Thus, as we outline below in
Section IV, we began with a model of methionine metabolism that had only
four variables (Reed et al., 2004). Then we made a model of the folate cycle
(Nijhout et al., 2004) so we could study the inhibition of DHFR by
methotrexate and the allosteric binding of folates to folate enzymes. Then
we made a larger model combining to two smaller ones so that we could
study the effects of the inhibition of MTHFR by SAM and the inhibition of
GNMT by 5mTHF (Nijhout et al., 2006). There has been a long discussion
in the literature or the role of the folate cycle in mitochondria (Appling,
1991; Christensen and MacKenzie, 2005) so to investigate these questions
we added compartmentation and the mitochondrial reactions (Nijhout
et al., 2007a). At each stage we had to make difficult (imperfect) decisions
about which variables and interactions to include in the models.
We note that this difficulty of knowing where to draw the boundaries is
also a difficulty for the interpretation of laboratory experiments or clinical
observations. In an experiment one changes the system by, say, knocking
out a gene or introducing a chemical that binds to a particular enzyme. One
then measures the changes in a few variables (the results of the experi-
ment) and then makes conclusions about how the system functions. Implic-
itly, one is assuming that everything else besides what one measures is the
same (or can be considered the same), that the knocked out gene did not
affect other genes or that the inhibitor has no other effects but the intended
one. The interpretation of the experiment typically involves implicitly
drawing the boundaries of a model (in the experimenters head) of
which variables are allowed to be included in the interpretation. Thus,
the interpretation of experimental results must always face and answer
(albeit implicitly) the same difficult question of boundaries faced explicitly
in modeling.
The next difficulty is deciding what level of detail to include for individ-
ual reactions in FOCM. An enormous amount of information is available
about enzymes, their genes and conformations and the way that they bind to
substrates. How much of this detail should be included? Our approach is to
use simple Michaelis-Menten kinetics and simple kinetic forms for activa-
tions and inhibitions unless we have good reason to believe that a more
detailed treatment is necessary for important biological functions of FOCM.
For example, we could have modeled the synthesis of SAMfrommethionine
in liver cells by a simple Michaelis-Menten formula. But we knew from the
56 H. F. Nijhout et al.
experiments of Finkelstein et al. (1982) and Finkelstein and Martin (1984,
1986) that the methionine levels are fairly stable under methionine loading
whereas SAM increases enormously, and that Corrales et al. (2002) had
suggested that this is a result of the different kinetics of the two isoforms
MAT-I and MAT-III. Because we believe that the stabilization of methio-
nine and the many regulations by SAM are biologically important, we
decided to include the rather complicated special kinetics of MAT-I and
MAT-III in the model. Subsequently, we were able to show (unpublished)
that the suggestion of Corrales et al. (2002) was completely correct.
Finally, of course, one must choose V
max
, K
m
, and K
i
values. This is not
such an easy matter since there are few measurements of V
max
values and the
reported measurements of V
max
, K
m
, and K
i
values show large variation.
Given the stochastic variation in gene expression discussed above and the
dependence of protein conformation and function on the in vivo context,
this variation is not surprising. We try to choose K
m
, and K
i
values within
reasonable experimental ranges and adjust the V
max
values so that the values
of the metabolite concentrations are in the experimental ranges. It is always
a question whether the results of in silico experiments would have been
different if we had chosen different parameters. We have found that most of
our qualitative results are quite insensitive to variations in parameter values.
In some sense, it has to be that way because FOCM must have evolved to be
able to continue to function in the face of the stochastic variation in
gene expression discussed above. Nevertheless, all three difficulties that
we have discussed necessarily temper the confidence that one has (that we
have!) in model results.
D. Advantages of mathematical models
Although models have difficulties and limitations, they also have advantages
and it is worthwhile to state them explicitly. First, to formulate a model one
has to be explicit about ones assumptions. If A inhibits B one must say how
much B is inhibited at different concentrations of A and how this may
or may not depend on other variables in the system. Secondly, once one has
a model, in silico experimentation is cheap, fast, and easy. One does not need
animals, IRB protocols, or technicians. Third, and most important, when
the model behaves in the same way as interesting experimental results, one
can take the model apart and put it back together (by removing reactions or
inhibitions, for example) until one understands the causal chain of events
that gives rise to the observed behavior. Thus, experiments with the model
can give real biological understanding of the phenomena under study.
Finally, in the model, one can follow the time course of all concentrations
and velocities and determine how the system reacts to outside influences
or changes in internal parameters. This is impossible to do by in vivo
experimentation.
Mathematical Models of Folate-Mediated One-Carbon Metabolism 57
Every experimental scientist is a modeler because every hypothesis is
based on a conceptual model of how a system ought to behave. A mathe-
matical model is simply a way of making a conceptual model explicit by
describing and connecting all the underlying knowledge and assumptions.
If a mathematical model does not reproduce the known behavior of a
system then, obviously, the model is wrong. But if the model is based on
all known data, then the ancillary conclusion is that the knowledge of the
system must be inadequate. Thus, a model can reveal the inadequacy of
current data or concepts. The model can then be used to test hypotheses
about what kind of additional (or different) information can yield the
correct behavior, and this can stimulate research to verify those predictions.
Another important use of a model is to test hypotheses about mechan-
isms that are difficult to study experimentally. We will give two examples
from FOCM. The first comes from a series of studies by Finkelstein and
Martin (1984, 1986) and Finkelstein (1990, 2001) who studied the allosteric
effect of SAM on the CBS and BHMT. They suggested that the concen-
tration of SAM rose with methionine input and that the allosteric stimula-
tion of CBS and inhibition of BHMT by SAMwould result in an increased
transsulfuration of homocysteine, which removes the excess methionine
from the system. Thus, the allosteric regulations by SAM constitute a
homeostatic mechanism that stabilizes the mass in the methionine cycle.
Our simulations with a model of the methionine cycle (Reed et al., 2004)
show that variation in methionine input is completely absorbed by variation
in the concentration of SAM. The model also shows that the allosteric
regulation of BHMT and CBS by SAMincreases the transsulfuration rate in
such a way that total mass in the methionine cycle, and the flux around the
methionine cycle, remain stable in the face fluctuating methionine input, as
first hypothesized by Finkelstein and Martin (1984).
The second example involves the role of the mitochondrial bifunctional
enzyme. In the mitochondria, the MTD MTCH reactions are catalyzed
by a single bifunctional enzyme (Mejia and MacKenzie, 1986; Peri et al.,
1989). This enzyme is not normally expressed in adult cells; it is expressed
only during embryonic development and in cancer cells (Di Pietro et al.,
2004; Smith et al., 1990), so its expression appears to be restricted to cells
undergoing high rates of cell division. On the basis of interpretation of a
series of radiotracer and gene knockout experiments, Christensen and
MacKenzie (2005) hypothesized that the bifunctional enzyme provides a
metabolic switch that controls the flow of one-carbon units to determine,
for example, the degree to which mitochondria produce formate and/or
convert glycine to serine. This hypothesis was confirmed by our mathe-
matical model (Nijhout et al., 2007a). Elimination of the mitochondrial
bifunctional enzyme in the model did not show a runaway accumulation of
CH
2
-THF, as might be expected. Instead, the GDC reaction slowed down,
the production and export of formate stopped entirely, and most
58 H. F. Nijhout et al.
importantly, the mitochondrial SHMT reaction reversed direction and now
ran toward serine synthesis. Thus, in the presence of the bifunctional
enzyme, a situation typical of embryonic and cancer cells, the mitochondria
export large quantities of formate that are directed to purine and TS in the
cytosol. When the bifunctional enzyme is not expressed, as in adult cells that
do not divide, the mitochondrial reactions become strong producers of
serine, which is exported to the cytosol and where it is directed toward
gluconeogenesis and other reactions. The bifunctional enzyme switch in
effect transforms the mitochondria from formate factories into serine fac-
tories, and may thus be an adaptation to the very different metabolic and
biosynthetic needs of rapidly growing embryonic cells and more quiescent
adult cells, as suggested by Christensen and MacKenzie (2005).
E. Kinetics, parameter values, and model structure
The reported diversity of parameter values for the same enzyme can be due
to various reasons: (1) the orthologous enzymes from different species can
have different kinetic properties; (2) enzyme expression differs in different
tissues, in particular some enzymes are up-or downregulated in cancers as
well as in tissues of animals undergoing chronic nutrient or vitamin depri-
vation or excess; (3) different semipurified enzyme preparations may con-
tain different, and unknown, concentrations of allosteric activators or
inhibitors; (4) enzyme preparations made at different times of day can
contain different concentrations of metabolites and allosteric effectors;
(5) in bimolecular reactions the values of the K
m
s depend on the concen-
tration of both substrates (Segel, 1975), but it is common to maintain one of
the substrates constant, resulting in the measurement of an apparent K
m
that
can differ depending on the preparation used.
Our approach to modeling the kinetics of one-carbon metabolism is to
restrict our use of reported kinetics to those measured in mammals, prefera-
bly humans, and we differentiate between parameters measured in different
tissues by building different models that specifically deal with hepatic
one-carbon metabolism and epithelial one-carbon metabolism (Figs. 2.3
and 2.5). Although measures of kinetic parameter values can vary signifi-
cantly, fortunately metabolite concentrations can be measured with great
accuracy and consistency, and the actual flux through a particular reaction,
or the relative dimensions of the fluxes through different portions of the
pathway are sometimes known. We start our modeling by choosing a value
of each K
m
and K
i
roughly in the middle of the reported range, and vary the
V
max
to obtain the reported metabolite concentrations and fluxes. We have
found experimental V
max
values to be largely not useful for modeling
purpose since they are typically reported in units of rate/mg protein,
without stating how protein was determined. We use values of the k
cat
in
those few cases where in vivo enzyme concentrations are known. We have
Mathematical Models of Folate-Mediated One-Carbon Metabolism 59
found that the choices of V
max
values are often constrained by the require-
ment that the model produce the right combination of known metabolite
concentrations, relative flux rates, half-lives, and time-dependent responses
to perturbations in the experimental literature.
When the kinetic mechanism of an enzyme is known we use the
conventional equations for the relevant uni- or bimolecular reaction as
described by Segel (1975). Allosteric activation or inhibition of enzymes
often does not admit to one of the conventional equations. In such cases, we
do a nonlinear regression on the published experimental data and use that as
the empirical equation. In these, as in all, cases, we ensure that the model
operates within the limits of the experimental data. Our models assume that
certain substrates (dUMP, GAR) and energetic metabolites (ATP, NADP,
and NADPH) are constant, so that their effect is absorbed by the V
max
for
the reaction.
At present the model does not contain terms for polyglutamation and
deglutamation. The model also does not contain a nuclear compartment,
although it is known that nuclear compartmentation is important (Appling,
1991; Woeller et al., 2007). We set the total folate level in the cell by
defining the overall size of the folate pool. If we start the simulation with all
folates in one form (e.g., THF or 5mTHF), the reaction kinetics rapidly
redistribute the folates, and the system comes to equilibrium for the differ-
ent folate species in 56 h. The half-life for folate in the body is about 90
days, and the mean residence time for folate is 124212 days (Gregory and
Quinlivan, 2002; Gregory et al., 1998), so for short-term studies like the
ones we do, the assumption of a constant intracellular folate pool seems
reasonable.
Glycine
5mTHF
THF
5,10-CH2-THF
S
H
M
T
Serine
DHF
10f-THF
TS
MS
DHFR
dUMP
dTMP
Pyrimidine
synthesis
5,10-CH=THF
MTHFR
MTD
MTCH
FTS
N
E
PGT
AICART
Purine
synthesis
AICAR
NADPH
NADPH
NADPH
NADP
+
NADP
+
NADP
+
GAR
HCOOH
FTD
SAM Methionine
SAH Hcy
SAHH
CBS
GNMT
MAT-I
MAT-III
DNA
DNA-CH
3
METin
Sarcosine
Adenosine
ATP
Betaine
BHMT
H
2
0
Glycine
DNMT
Methylation
reactions
H
2
C=O
Folate cycle Methionine cycle
Serine
Cystathionine
Figure 2.3 Diagram of our model of the coupled hepatic folate and methionine cycles.
Not shown in this figure are all the allosteric interactions between metabolites and
various enzymes in the two cycles. Full names of the acronyms of enzymes and
metabolites for this and other figures, and the text, are given in Table 2.1. Redrawn
from Reed et al. (2006).
60 H. F. Nijhout et al.
The model then, consists of a set of kinetic formulas, one for each
enzyme, that describe the velocity of the reaction as a function of the
concentrations of substrates, products and allosteric regulators, plus a set of
differential equations, one for each variable metabolite, that contain the
kinetic formulas for its synthesis and degradation. In addition, we have
transport functions for amino acids into and out of the cell, and or amino
acids and formate into and out of the mitochondria. The overall system is
solved by numerical integration using a stiff ode solver (because different
quantities tend to vary at very different rates), implemented in MatLab
(The MathWorks). The program allows us to vary inputs of amino acids
and vitamins over time and follow the time-dependent responses of all
metabolites and reaction rates. In addition, the model allows us to simulate
the effects of mutations and of vitamin deficiency (or excess). We have
modeled mutations primarily by altering the V
max
values of the relevant
enzymes. This would correspond to mutations that affect the amount of
active enzyme present (e.g., mutations that affect enzyme expression or
activation). Likewise, we model the effects of variation in non-folate vita-
min cofactors, such as B
12
and B
6
, by altering the V
max
of the corresponding
enzyme(s), assuming in effect that the activity of the enzyme is a function of
the amount of cofactor available.
IV. Model Development
Previous models of folate metabolism, outlined above, were devel-
oped in the 1970s and 1980s. Much new information and understanding
have become available in the intervening 25 years, which have guided our
approach. We began by first developing a model for the methionine cycle
(Reed et al., 2004). This model built on the prior work of Martinov et al.
(2000) who had studied the properties of a model for a portion of the
methionine cycle that did not include the MS, BHMT, and CBS reactions
and used simplifying assumptions about inputs into the cycle. Our model
closed the cycle and added the CBS reaction and several allosteric effects of
SAM. This model was able to reproduce the observed dependence of the
transsulfuration reaction on the concentration of SAM described by
Finkelstein and Martin (1984), and the effects of variation in CBS and MS
activity on homocysteine, methionine, and SAM (Finkelstein, 1990;
Finkelstein et al., 1974; Janosik et al., 2001; Pogribna et al., 2001;
Rosenblatt, 2001). Perhaps the most interesting finding with this model
was that SAM acts as a buffer for methionine input: that is, variation in
methionine input has little effect on the methionine and homocysteine
concentrations but is mostly absorbed by variation in the concentration of
SAM. Furthermore the allosteric effect of SAM on CBS provides a
Mathematical Models of Folate-Mediated One-Carbon Metabolism 61
mechanism for stabilizing mass in the methionine cycle so that the flux out
of the methionine cycle via CBS matches the rate of methionine input into
the cycle without much change in the homocysteine concentration. If it
were not for the allosteric effect of SAM, the homocysteine concentration
would have to rise to drive the CBS reaction.
Our next step was to develop a model for the folate cycle that
contained what at the time we understood to be the important reactions
of that metabolic network (Nijhout et al., 2004). This model incorporated
the finding that folates bind to and inhibit many of the enzymes in the
folate cycle. This binding was believed to provide a reservoir of folates. The
model allowed us to resolve the puzzle as to why enzymes of the folate
cycle should be inhibited by allosteric binding of folates. The model shows
that this nonenzymatic binding greatly reduces the sensitivity of the system
to folate deficiency, because as the total pool of folate diminishes, more
enzyme is released from inhibition, and the reaction velocities are main-
tained because of the increased enzyme activity (Nijhout et al., 2004).
We next modeled the allosteric interactions between the folate and methi-
onine cycles (Fig. 2.11) in order to test the hypothesis of Wagner et al.
(1985) that these interactions serve to stabilize the DNA methylation
reaction rates (Nijhout et al., 2006). Some results of these experiments are
outlined in Section VII below. We subsequently merged our models for the
folate and methionine cycles (Fig. 2.3) to produce an integrated model of
one-carbon metabolism (Reed et al., 2006). This model also incorporated
allosteric interactions between the folate and methionine cycles (inhibition
of MTHFR by SAM and SAH, and inhibition of GNMT by 5mTHF) and
added the ability to vary the rate of input of betaine. We used this model to
simulate the interaction between folate deficiency and the MTHFR C677T
polymorphism and the interaction between folate and vitamin B
12
deficien-
cies. Experimentation with this model showed that the inverse relationship
between folate status and homocysteine level is strongest at low folate levels
and disappears at high folate levels. Furthermore, the model shows that as
folate levels in the cell rise, the reactions of the folate cycle slow down. This
is due to the allosteric inhibition of enzymes in the folate cycle by folate
metabolites. This is a consequence of the homeostatic mechanism described
by Nijhout et al. (2004). This mechanism stabilizes the folate cycle at low
and intermediate folate levels, but also predicts that as folate levels rise, the
reaction rates in the folate cycle will slow down. Thus, a prediction of the
model is that a high intracellular folate level can have the same effect as a
folate deficiency. This prediction of the model is now supported by a variety
of clinical and experimental data that show that high doses of folate can have
detrimental effects (Akoglu et al., 2001; Czeizel, 2004; Morris et al., 2005;
Sunder-Plassman et al., 2000; Troen et al., 2006).
We then expanded the model of Reed et al. (2006) to include com-
partmentation of the folate cycle between cytosol and mitochondria
62 H. F. Nijhout et al.
5mTHF
THF
5,10-CH2-THF
Serine
DHF
10f-THF
MS
TS
dUMP
Pyrimidine
synthesis
dTMP
5,10-CH=THF
MTD
MTCH
FTS
NE
PGT
AICART
Purine
synthesis
NADPH
DHFR
NADPH
NADP
+
NADP
+
MTHFR
NADP
+
NADPH
GAR
HCOOH
FTD
SAHH
CBS
GNMT
MAT-I
DNA
DNA-CH
3
Methionine
Homocysteine SAH
SAM
Metin
Adenosine
ATP
Betaine
BHMT
MAT-III
H
2
O
DNMT
H
2
C=O
Cystathionine
THF
5,10-CH2-THF
10f-THF
5,10-CH=THF
MTD
MTCH
FTS
NE
NADPH
NADP
+
HCOOH
FTD
H
2
C=O
Glycine
Mitochondria
Cytosol
GDC
Methionine
5mTHF
CO
2
SDH
DMGD
CO
2
CO
2
Ser Ser
Gly
Ser
Gly
Gly Gly
Gly
DMG
Sarc
Sarc
DMG
Sarc
Gluconeogenesis
AICAR
cSHMT
mSHMT
Methylation
reactions
Figure 2.4 Diagram of our model of hepatic FOCM including the mitochondrial compartmentation. Boxed substrates are variables in the
model. Substrates without boxes are constants. Full names of the acronyms of enzymes and metabolites for this and other figures, and the
text, are given in Table 2.1. Redrawn from Nijhout et al. (2007a).
(Nijhout et al., 2007a). This model included terms for the glycine cleavage
system and the metabolism of sarcosine and dimethylglycine in the mito-
chondria, mechanisms for transport of serine and glycine into the cell and
between the cytosol and mitochondria, and terms for the transport of
formate between cytosol and mitochondria (Fig. 2.4). As discussed above,
we discovered that in rapidly dividing cells mitochondria act primarily to
supply formate to the cytosol for purine and pyrimidine synthesis, whereas
in adult cells the mitochondria export no formate but are excess producers
of serine, targeted for gluconeogenesis. We also found that the rate of export
of formate fromthe mitochondria to the cytosol is remarkably insensitive to
fluctuations in serine and glycine input. This is because both mitochondrial
and cytosolic SHMT reactions are reversible and the rates at which they run
are highly responsive to the relative concentrations of glycine and serine.
The model was used to investigate the effect of varying the relative inputs of
glycine and serine on the rate and direction of the mitochondrial and
cytosolic SHMT reactions, and showed that both SHMT reactions can
reverse and run in the serine synthesis direction when external glycine is
increased replicating the results of Kastanos et al. (1997). This model was
also used to successfully simulate the experiments of MacFarlane et al. (2005)
and Herbig et al. (2002) on the effect of SHMT expression and glycine
availability on SAM.
To investigate the characteristics of FOCM in nonhepatic tissues we
developed a model for epithelial FOCM, which is representative of most
tissues except liver and kidney. Extrahepatic tissues do not express all
enzymes of FOCM, and some enzymes are active at much lower levels
than in the liver (dashed arrows in Fig. 2.5 ). Epithelia thus run on a reduced
version of the network. This model also includes a term for export of
homocysteine, which is typically exported from extrahepatic tissues for
remethylation in the liver. With this model we have explored the interac-
tion of multiple genetic polymorphisms and the interaction of genetic
and environmental variation on the level of homocysteine, the rates of
methylation, and purine and pyrimidine synthesis (Ulrich et al., 2008).
We have also created a model (Fig. 2.1) that includes the synthesis of
reduced glutathione and exchange of substrates with the blood (Reed
et al., 2008).
FOCM is an ancient pathway and occurs, with variations, in animals,
plants, fungi, and bacteria. Recently, we have developed a model of bacte-
rial FOCMfor Rhodobacter capsulatus (Leduc et al., 2007), motivated by the
discovery of a novel flavin-dependent thymidylate synthase (ThyX) that
produces THF rather than DHF upon methylation of dUMP (Fig. 2.6).
This model was used to examine the relative roles of ThyA (TS), ThyX, and
FolA (DHFR) in the mechanism of resistance to antifolates such as
trimethoprim.
64 H. F. Nijhout et al.
Glycine
5mTHF
THF
5,10-CH2-THF
S
H
M
T
Serine
DHF
10f-THF
TS
MS
DHFR
dUMP
dTMP
Pyrimidine
synthesis
5,10-CH=THF
MTHFR
MTD
MTCH
FTS
N
E
PGT
AICART
Purine
synthesis
AICAR
NADP
+
NADPH
NADPH NADP
+
NADP
+
NADPH
GAR
HCOOH
H
2
C=O
FTD
SAM Methionine
SAH Hcy
SAHH
CBS
MAT-II
DNA
DNA-CH
3
METin
Adenosine
ATP
H
2
O
DNMT
Methylation
reactions
Folate cycle Methionine cycle
Serine
Cystathionine
METH
Figure 2.5 Diagram of our mode of epithelial FOCM. This model does not include mitochondrial compartmentation, but does include all
allosteric regulations (not shown in this diagram). Dashed arrows indicate enzymes of low activity in epithelial tissues. Full names of the
acronyms of enzymes and metabolites for this and other figures, and the text, are given in Table 2.1.
V. Blood Versus Intracellular Metabolite
Concentrations
The models we have developed are for intracellular metabolism, and
thus deal with intracellular concentration and pool sizes. However, almost all
of our understanding of the relationships between folate status and disease is
based on measurements of the concentrations of folate, homocysteine, SAM,
SAH, and methionine in the blood, plasma, or red blood cells. Red blood
cell measurements are believed to reflect the metabolite status at the time the
red blood cells differentiated: in a mixed-age population of cells this pre-
sumably represents an average or long-term metabolic status. Blood and
plasma concentrations may be in equilibrium with overall cellular cytosolic
concentrations, though it is more likely that they result from the interaction
between uptake from the digestive system, export by some tissues (like
epithelial cells, kidneys, muscle, and nervous system), import by others
(like the liver), and excretion by the kidneys. Whether these processes are
ever a steady state is an open question. The half-life of folate in the body is
about 90 days, and about 500 days are required for folate levels to come to a
new steady state (Gregory et al., 1998). Methionine loading experiments
shows that the methionine and homocysteine levels in the blood require
1224 h to return to steady state after a perturbation (Bianchi et al., 2000;
Glycine
5mTHF
THF
5,10-CH2-THF
Serine
DHF
10f-THF
TS
ThyA
MS
MetH
DHFR
FolA
dUMP
dTMP
Pyrimidine
synthesis
5,10-CH=THF
MTHFR
MetF
MTD
FolD
MTCH
FolD
PGT
PurN
AICART
PurH
Purine
synthesis
AICAR
GAR
HCOOH
FTS
PurU
Methionine
Hcy
Glycine
Ammonia+CO
2
ThyX
dTMP
dUMP
Pyrimidine
synthesis
5f-THF
SHMT
GlyA
GDC
GCVT
MTS
DHPR
DHFR
FolA
DHFS
FolC
Folate
Dihydropteroate
SHMT
GlyA
Figure 2.6 Diagram of our model of bacterial folate metabolism. This mechanism can
synthesize thymidylate but bypass DHF. Redrawn from Leduc et al. (2007).
66 H. F. Nijhout et al.
Silberberg and Dudman, 2001). In our models, the time required for differ-
ent components of the system to relax to equilibrium is in the order of hours
to days (Fig. 2.7). Given that variation in input into the systemis in the order
of hours, it is unlikely that the system is ever at steady state and may actually
exist far from equilibrium most of the time (Nijhout et al., 2007b).
Thus, blood measurements represent some average of what is going on in
different cell types, and one would therefore expect a variable and context-
dependent correlation between blood components (particularly for meta-
bolites that are used in many processes) and the state of a given organ or
cellular metabolic system. Our current intracellular models accurately simu-
late intracellular responses to experimental or clinical intervention, and it is
obviously desirable for a model to also simulate how the levels of commonly
measured blood metabolites will respond. We are beginning to approach this
difficult question of whole body modeling of folate metabolism by allowing
our cytosolic models for hepatic and epithelial FOCMto communicate with
a blood compartment that is subject to dietary input and excretory output.
VI. Modeling GeneGene and GeneEnvironment
Interactions
One advantage of models is that they can be used to investigate the
effects of simultaneous variation of many variables. In the case of FOCM,
the models can be used to study the effect of simultaneous genetic
8000
7500
7000
6500
6000
5500
5000
4500
4000
50 40
Time (h)
C
o
n
c
e
n
t
r
a
t
i
o
n
s
a
n
d
v
e
l
o
c
i
t
i
e
s
[
G
S
H
+
G
S
S
G
]
m
M
vGSHout
[SAM]
[cCys]
[bCys]
vCBS
[GSH+GSSG]
30 20 10 0
0
50
150
200
100
Figure 2.7 Response of selected metabolite concentration and reaction velocities to
fasting. The system was allowed to come to steady state, and at 3 h the input of amino
acids (glycine, serine, methionine, andcysteine) was reducedto0.25of their normal values.
Different components of FOCM decline at different rates to a new steady state. Some
approachsteadystateafter 56 h, other takemorethan48 htoapproachthenewsteadystate.
Mathematical Models of Folate-Mediated One-Carbon Metabolism 67
polymorphisms, or the interaction between a polymorphism and an envi-
ronmental variable such as an amino acid, vitamin B
12
, or folate. For
instance, the interaction of the MTHFR C677T polymorphism with low
folate status is shown in Fig. 2.8. The T/T genotype is known to diminish
the risk of colon cancer under high folate, but it enhances risk for cardio-
vascular disease (Curtin et al., 2004; Frosst et al., 1995). In the model, the T/
T genotype lowers the concentration of SAM and the DNMT reaction rate
but raises homocysteine levels and both thymidylate and purine synthesis
rates. Folate deficiency enhances the effects of the T allele on most biomar-
kers, with the exception that it reverses the effect on thymidylate and purine
synthesis. Although these simulated changes in biomarkers correspond to
those observed in practice, it is not yet clear how these metabolic effects
translate into differential risks for colon cancer and cardiovascular disease.
The interaction between variation in the V
max
of MS and of MTHFR is
illustrated in Fig. 2.9. The variation along the MS axis in Fig. 2.9 can be
interpreted in several ways. It can represent variation in the expression level
of MS which could be due to a regulatory mutation (e.g., in the promoter
region of the MS gene), or it could be due to mutations in a structural gene
that affects the k
cat
. Variation could also be due to variation in the level of
vitamin B
12
, which is a cofactor for MS. In the first two cases variation is
genetic, and in the latter case the variation is environmental (e.g., due to a
vitamin B
12
deficiency). Allowing parameters to vary continuously makes it
possible to explore a broad range of possibilities (corresponding to a range of
alleles with minor effects, or a range of environmental exposures) around
the normal or wild type, indicated by the open circles in Fig. 2.9. The shape
100
80
60
40
20
0
20
40
60
R
e
l
a
t
i
v
e
c
o
n
c
e
n
t
r
a
t
i
o
n
o
r
r
a
t
e
C/C
C/T
T/T
T/T (low folate)
DNMT
rate
[SAM]
[SAH]
[Hcy]
[5mTHF]
Thymidylate
synthesis Purine
synthesis
Figure 2.8 Simulated response of various biomarkers to the MTHFR 677CT polymor-
phism. Folate deficiency exacerbates the effect of the T/T genotype for most biomarkers,
but reverses the effect of the T/T genotype on purine and pyrimidine synthesis.
68 H. F. Nijhout et al.
of the interaction relationship is clearly nonlinear, so the effect of variation
in MS depends on the exact value of MTHFR activity, and vice versa.
The effects of the interaction of MS and SHMT activity on purine
synthesis and homocysteine concentration are shown in Fig. 2.9. These
relationships are also nonlinear, as indeed are all relationships between
variables within FOCM. In many cases, such as the ones illustrated here,
the wild-type values lie on a relatively flat and horizontal region of the
phenotypic surface. This indicates that the wild type is relatively insensitive
to variation in parameter values, because modest variation of the variables or
parameter values (x and y axes) has little effect on the phenotype (z axis).
As the parameter values move far away from the normal, or wild type, the
effect of their variation increases dramatically.
The finding that many wild-type phenotypes lie in regions of the
phenotypic surface that are relatively flat and horizontal, implies that the
Figure 2.9 Bivariate graphs showing the interaction of various enzymes in FOCM on
selected traits (Z axes). X and Y axes show variation in the V
max
of the respective
enzymes. This variation could be due to allelic variation or, in the case of MS, to variation
in vitamin B
12
. The circle shows the location of the normal or wild-type phenotype.
Mathematical Models of Folate-Mediated One-Carbon Metabolism 69
system is relatively insensitive to the exact values of those parameters. From
an evolutionary perspective one would therefore not expect strong selection
to maintain those parameter values within close tolerances, because moder-
ate variation has little effect on the phenotype. This observation may help
explain why reports on parameter values from different preparations are
often inconsistent. Although differences in protocols and experimental
errors surely play a role, it is not unreasonable to assume that some of this
variation may be real. It is possible that many genes accumulate small-effect
mutations in their regulatory region, or their coding region, that would be
neutral to selection. Indeed, human genes exhibit abundant single nucleo-
tide polymorphism (SNP) variation. The HapMap Project has uncovered a
polymorphic SNP on average every 825 base pairs, and on the average
2 nonsynonymous SNPs per gene (International HapMap Consortium,
2007; McVean et al., 2005). In addition, there is a large amount of regu-
latory variation in the promoter region of genes that leads to variation in the
level of expression (Rockman and Wray, 2002; Yan et al., 2002). A survey
of naturally occurring polymorphisms in the promoter regions of 107
human genes showed that 60% caused more than a twofold difference in
expression, and 11% caused more than a tenfold difference in expression
(Rockman and Wray, 2002). Finally, within a genetically identical popula-
tion of cells the concentration of a given protein can vary by as much as 30%
fromcell to cell and fromtime to time (Elowitz et al., 2002; Sigal et al., 2006).
Thus, there is far more individual genetic variation and individual
variation in gene expression than is typically assumed. This, together with
the fact that the metabolites and allosteric effectors involved in FOCM vary
among individuals and from time to time (e.g., Fig. 2.10), suggests that
much of this variation is without significant effect on fitness, and is therefore
not under selection, and may therefore explain some of the observed
interindividual variability.
VII. Modeling and Simulation have Revealed
Novel Homeostatic Mechanisms
FOCM has many functions that must continue to operate normally in
face of variation in the demand on specific reactions and variation in the
input of metabolites. For instance, the expression levels of TS and DHFR
are upregulated more than 100-fold during the S-phase of the cell cycle,
when there is an increased demand for nucleotide synthesis (Bjarnason et al.,
2001; Obama et al., 2002; Slansky et al., 1993; Wade et al., 1995). At the
same time there will be an increased demand for DNA methylation to
maintain the correct methylation pattern of the newly synthesized DNA
strands. FOCM is also subject to great hourly and daily variation in amino
acid input, which varies with meals and nutrition. The amino acids serine
70 H. F. Nijhout et al.
160
A
B
C
140
120
100
80
60
200
150
100
50
0
50
100
200
150
100
50
0
0 5 10 15 20
Time (h)
C
o
n
c
e
n
t
r
a
t
i
o
n
s
(
m
M
)
R
e
a
c
t
i
o
n
r
a
t
e
s
(
m
M
/
h
)
TS
mSHMT
[Methionine]
[SAM]
100*[Hcy]
cSHMT
CBS
AICART
HCOOH/10
DNMT
Figure 2.10 Simulated response of selected hepatic FOCM metabolite concentrations
and reaction velocities to periodic pulses of amino acid input. (A) Concentration
profiles of methionine cycle metabolites. (B) Velocity profiles of the CBS reaction
and the mitochondrial and cytosolic SHMT reactions (mSHMT and cSHMT, respec-
tively). (C) Velocity profiles of the DNMT, TS, and AICART reactions and the rate of
export of formate (vHCOOH) from the mitochondria. Modified from Nijhout et al.
(2007b).
Mathematical Models of Folate-Mediated One-Carbon Metabolism 71
and glycine are the primary methyl donors for FOCM, and methionine is
both a methyl donor and an essential amino acid linking the folate and
methionine cycles. An interesting question is whether and how the stability
of critical reactions in the cycle are maintained when there are large
localized changes in demand, or large localized changes in input.
Perhaps the best way to illustrate the relative stability of some reactions
in the face of variation in inputs is by simulating a day in the life of FOCM.
After each meal with protein, the human body experiences a pulse of amino
acids that lasts about 3 h. We simulated this by pulsing the four amino acids
that serve as inputs for the model (Fig. 2.10). It is evident from the
simulations shown in Fig. 2.10 that some variables change dramatically
with each meal, while others are almost unaffected. The TS and DHFR
reactions are quite stable as are the DNMT rate and the rate of export of
formate from the mitochondria. By contrast, the SHMT reactions fluctuate
greatly as do the concentrations of SAM and homocysteine.
As discussed above (Section III.D), the stability of formate export from
the mitochondria arises from the dynamical interplay between the mito-
chondrial and cytosolic SHMT reactions, whose magnitude and direction
vary with serine and glycine input. The fluctuations in SHMT velocity are a
dynamic homeostatic mechanism that dampens the effects of fluctuations in
glycine and serine input (Nijhout et al., 2007a).
The methylation of DNAis an important function of FOCM, and it seems
reasonable to stabilize these reactions against a variable and often unpredictable
input of methyl groups. Simulations with our models show that the DNMT
reaction is extraordinarily stable against variation in input, and that this stability
arises from two allosteric interactions between the folate cycle and the methi-
onine cycle: SAM inhibits MTHFR and 5mTHF inhibits GNMT (see
Fig. 2.11). Wagner et al. (1985), and Wagner (1995) suggested that the purpose
of these interactions might be to stabilize the rate of DNA methylation. The
general idea of how this mechanism works is easy to understand. If the
concentration of SAM goes up, then MTHFR is inhibited, which causes the
concentration of 5mTHF to fall. When 5mTHF is lower, the inhibition of
GNMT is released causing the rate of the GNMT reaction to go up, utilizing
the extra SAM and allowing the DNMT rate to remain stable. The reverse
scenario explains what happens if SAM goes down.
We experimented with the model shown in Fig. 2.11 by adding and
removing the long-range allosteric regulations in various combinations.
Figure 2.12 shows how the [SAM]/[SAH] ratio and the DNMT reaction
rate vary as a function of methionine input under two scenarios: with all
allosteric interaction present, and with all allosteric interactions absent. It is
clear that the allosteric interactions stabilize the SAM/SAH ratio and the
DNMT reaction rate against variation in methionine input, and that the
effect is most pronounced at low methionine input. This is because under an
optimal supply of methionine the DNMT reaction runs close to saturation,
72 H. F. Nijhout et al.
so the main benefit of these regulations appears to be to protect the DNMT
reaction against periods of protein starvation.
Finally, as noted above, the expression of TS and DHFR vary a 100-fold
or more with various stages of the cell cycle, and we have shown, by
simulation, that this variation has little or no effect on the reaction velocities
and metabolite concentrations elsewhere in the folate and methionine
cycles (Nijhout et al., 2004). An implication of this funding is that FOCM
should be relatively insensitive to inhibition of TS and DHFR by chemo-
therapeutic drugs such as methotrexate (which inhibits DHFR) and
5-fluoro-uracil (which inhibits TS). This is indeed the case. When the
V
max
of DHFR is lowered (corresponding, e.g., to treatment with metho-
trexate) the velocity of the DHFR reaction remains virtually constant until
there is almost no free enzyme left. The reason for this remarkable stability
of the DHFR reaction is that the normal concentration of its substrate,
DHF, is exceptionally low, typically 0.02 mM out of a total folate pool of
20 mM. Thus, the concentration of DHF can rise more than a 100-fold
THF
NADP
+
NADPH
5mTHF
MS
MTHFR
Cystathionine
DNA
DNA-CH
3
METin
Sarcosine
Adenosine
ATP
Betaine
H
2
O
Glycine
Methionine
SAH Homocysteine
SAHH
CBS
BHMT
MAT-I
MAT-III
DNMT GNMT
5,10-CH
2
TFH
(folate cycle)
Serine
SAM
Figure 2.11 Model of the allosteric regulatory interactions within the methionine
cycle and between the folate and methionine cycles used to study the stability of the
DNMT reaction against fluctuations in methionine input. Thick lines show the alloste-
ric interactions of SAM and 5mTHF. Arrow indicates activation and bars indicate
inhibition. Redrawn from Nijhout et al. (2006).
Mathematical Models of Folate-Mediated One-Carbon Metabolism 73
(and drive the DHFR reaction by substrate accumulation) without substan-
tially depleting the folate pool and disrupting the reaction rates elsewhere in
FOCM. The inhibition of DHFR by methotrexate in effect creates a DHF
160
A
B
140
120
100
80
60
40
20
20
15
10
5
0
0 20 40 60 80 100 120 140
Methionine input (mM/h)
0 20 40 60 80 100 120 140
Methionine input (mM/h)
Unregulated
Regulated
Unregulated
Regulated
S
A
M
/
S
A
H
r
a
t
i
o
M
e
t
h
y
l
a
t
i
o
n
r
a
t
e
(
m
M
/
h
)
Figure 2.12 Effect of allosteric regulation by SAM and 5mTHF on the response of
(A) the [SAM]/[SAH] ratio and (B) the DNMT reaction to variation in methionine
input. Solid line shows the response when all allosteric regulations are in place, and the
dotted line shows the response without allosteric regulation. Allosteric regulation
stabilizes both the [SAM]/[SAH] ration and DNMT reaction at low methionine
input, and thus may be an adaptation to protein starvation.
74 H. F. Nijhout et al.
trap. The rate at which this DHF trap develops is determined by the rate of
the TS reaction. In a rapidly dividing cancer cell, where TS is highly up-
regulated, the DHF trap will develop rapidly, whereas in a non-cancerous,
non-dividing cell it will develop very slowly if at all.
VIII. Steady States and Fluctuations
The mathematical models allow us to calculate how long it takes for
FOCM to return to steady state after a perturbation. The interlocking cycles
of FOCM are complex and different reactions return to steady state at
different rates. An example is shown in Fig. 2.7 where we show the
simulated response to fasting. It takes 610 h for some reactions to go to
steady state while others take more than 2 days. Given a normal pattern of
eating, these findings imply that FOCM is never at steady state (Nijhout
et al., 2007b). Indeed many reaction rates and metabolite concentrations are
likely to always be far from steady state (Fig. 2.10).
This calls into question the utility of standard metabolic control analysis
to understand the operation of this system. In metabolic control analysis one
typically lets the system come to steady state, then perturbs it by changing
one parameter by a small amount, and lets the system come to the new
steady state (Fell, 1992). The fractional change in the reaction velocities and
metabolite concentrations at this new steady state is then taken to be a
measure of the sensitivity of each component of the system to the parameter
that was changed. This method is used to deduce how control is distributed
among the reactions of a system, and the relative control any given enzyme
has over the operation of the system. Metabolic control analysis is, in effect,
a sensitivity analysis preformed by perturbing the steady state. When a
system normally operates far from steady state, and its reaction velocities
and metabolite concentrations are continually changing, a steady-state
sensitivity analysis is not a useful way of obtaining insight into the operation
of the system.
Instead, it is more natural to see how the system responds to large scale
fluctuations. We have been using such fluctuations in different ways. First,
we have been using fluctuations to make quantitative statements about the
effects of particular homeostatic regulatory mechanisms. For example, in
Nijhout et al. (2006) we added to the normal methionine input (100 mM/h)
a continuous stochastic fluctuation with standard deviation 30 mM/h. The
standard deviation in the velocity of the DNA methylation reaction was
exceptionally small, because of the long-range allosteric interactions dis-
cussed above. We then removed allosteric interactions one by one to see
which ones and which combinations had the greatest effects. When all four
are removed, the standard deviation of the velocity of the DNMT reaction
Mathematical Models of Folate-Mediated One-Carbon Metabolism 75
goes up by a large factor. We have also used such fluctuation analysis to
show that it is the unusual kinetics of MAT-I and MAT-III that stabilizes
the methionine concentration at the expense of large fluctuations in SAM
(unpublished). In Nijhout et al. (2007), we applied stochastic fluctuations to
the serine and glycine inputs and showed that the production of formate by
the mitochondria remains remarkably stable. This stability is caused by the
parallel SHMT reactions in the cytosol and the mitochondria that make
glycine from serine and vice versa.
Secondly, we often use external stochastic fluctuations as a probe of
system behavior. It is very interesting to fluctuate an input or a V
max
(corresponding to gene up- and downregulation) and then observe which
concentrations and velocities fluctuate a lot, a moderate amount, or hardly
at all. Our experience is that when a concentration or velocity hardly
fluctuates at all, there is usually a good biological reason why this is so.
We can then take the systemapart to discover the mechanisms that cause the
homeostatic behavior. Usually, some other concentrations and velocities
change a lot so that the homeostatic ones can remain stable.
Finally, we have been conducting a mathematical analysis of the way
in which general fluctuations propagate through biochemical networks.
In Anderson et al. (2007), we showed that the variances of reactions
velocities are always strictly decreasing down linear chains. The biological
significance of this result is that if it is important to stabilize the output of a
chain of biochemical reactions against fluctuations in the input, then the
chain should be long. It was also shown that side reaction systems and
feedback loops decrease the variations of the velocities in downstream
reactions. In Anderson and Mattingly (2008), many of these results are
proven in the case of Michaelis-Menten chains. Efforts are underway to
prove how more complicated network geometries and different kinds of
kinetics affect the ways in which fluctuations propagate.
Finally, we note that metabolic networks do not arise fully formed. They
evolve over time by the addition and elimination of reactions and by
changes in the kinetics of existing reactions. In evolutionary biology, it is
typically assumed that natural selection acts to maximize flux through a
pathway (e.g., Hartl et al., 1985; Wagner, 2005), in effect making reactions
more efficient in some way. But if a system normally experiences con-
tinual and large fluctuations of input, and continuous and large changes in
the demand for many different synthetic reactions, then a more likely target
for natural selection would be those reactions or connections that stabilize
certain parts of the system against the effects of those changes. That is,
evolution would not necessarily favor faster and more efficient pathways,
but rather would favor pathways that operate stably and reliably under
variation. Eating imposes enormous hourly and daily fluctuations, as well
as unpredictable long-term deficiencies in specific nutrients, and normal
daily and seasonal activities impose large variation in demand. This is true of
76 H. F. Nijhout et al.
FOCM, and it must be true of most if not all of metabolism. The key
regulatory features of metabolic systems are thus those that stabilize func-
tion, and those that prevent local perturbations from propagating through
the system. As is the case in FOCM, these regulatory mechanisms are not
the emergent properties of large networks, but are evolved adaptations for
specific functions.
IX. Conclusions
FOCM is one of the best studied metabolic systems: all or almost all
enzymes and metabolites in the system are known, as is the structure of
the reaction network. This network is complex and consists of several
intersecting cycles and a large number of complex allosteric regulatory
interactions between metabolites and enzymes. The reactions in this system
are nonlinear, which makes it exceptionally difficult to deduce the proper-
ties of the overall system, the way it is regulated, and the effects of mutations
and nutrient and vitamin deficiencies from the connectivity diagram alone.
The most direct way to understand the function of different parts of a
complex system like FOCM is through computer simulation with a mathe-
matical model. Because FOCM has been so well studied, it has been
possible to construct models that accurately simulate metabolite pools and
reaction velocities, as well as the effects of mutations and vitamin deficien-
cies on markers like homocysteine, TS and methylation capacities.
A mathematical model is an experimental tool that can be used as a
complement to laboratory experimentation or clinical investigation to do
pilot experiments and test hypotheses quickly and inexpensively. When a
new interaction is discovered, or suspected, it can be incorporated into a
preexisting model to determine its effect. We expect that our mathematical
models will evolve in three ways: first by progressive improvement of the
accuracy of the existing models by incorporating details like polyglutamation,
substrate channeling, and compartmentalization; second, by extending the
models to include other related aspects of metabolism, like insulin signaling;
third, by developing additional tissue-specific models, for instance, for the brain
and transport across the bloodbrain barrier, and by linking models for multiple
organ systems together through the circulatory system.
One important purpose of studying FOCM is to understand the rela-
tionship between genetic and environmental variables and disease out-
comes. There are two large steps necessary for this understanding. First,
one needs to understand how genetic and the environmental perturbations
affect the system behavior of FOCM. Second, one needs to understand how
the system changes in FOCM lead to the various disease states. Both are
very difficult questions. Most of our work outlined above has been
Mathematical Models of Folate-Mediated One-Carbon Metabolism 77
dedicated to understanding the regulatory system properties of FOCM and
how the behavior of FOCM changes in the presence of genetic polymorph-
isms and changes in environmental input. It remains a formidable challenge
to understand the pathway by which inadequacies or malfunctions of the
processes regulated by FOCM contribute to the development of such
diverse diseases as colon cancer, psychiatric disorders, cardiovascular disease,
and neural tube defects.
ACKNOWLEDGMENTS
We thank Marian Neuhouser, Jess Gregory, Barry Shane, Jill James, and Jon Mattingly for
their advice during the development of the mathematical models of FOCM. This work was
supported by grant DMS-0616710 from the National Science Foundation, and grant RO1
CA 105437 from the National Institutes of Health.
REFERENCES
Akoglu, B., Faust, D., Milovic, V., and Stein, J. (2001). Folate and chemoprevention of
colorectal cancer: Is 5-methyltetrahydrofolate an active antiproliferative agent in folate-
treated colon-cancer cells? Nutrition 17, 652653.
Anderson, D. F., and Mattingly, J. C. (2008). Propagation of fluctuations in biochemical
systems, II: Nonlinear chains. IET Syst. Biol. 1, 313325.
Anderson, D. F., Mattingly, J. C., Nijhout, H. F., and Reed, M. C. (2007). Propagation of
fluctuations in biochemical systems, I: Linear SSC networks. Bull. Math. Biol. 69,
17911813.
Appling, D. (1991). Compartmentation of folate-mediated one-carbon metabolism in
eukaryotes. FASEB J. 5, 26452651.
Bianchi, G., Brizi, M., Rossi, B., Ronchi, M., Grossi, G., and Marchesini, G. (2000).
Synthesis of glutathione in response to methionine load in control subjects and in patients
with cirrhosis. Metabolism 49, 14341439.
Bjarnason, G. A., Jordan, R. C., Wood, P. A., Li, Q., Lincoln, D. W., Sothern, R. B.,
Hrushesky, W. J. M., and Ben-David, Y. (2001). Circadian expression of clock genes in
human oral mucosa and skin association with specific cell-cycle phases. Am. J. Pathol.
158, 17931801.
Christensen, K. E., and MacKenzie, R. E. (2005). Mitochondrial one-carbon metabolism is
adapted to the specific needs of yeast, plants and animals. Bioessays 28, 595605.
Clarke, S., and Banfield, K. (2001). S-Adenosylmethionine-dependent methyltransferases.
In Homocysteine in Health and Disease (R. Carmel and D. W. Jacobsen, eds.),
pp. 6378. Cambridge Univ. Press, Cambridge.
Cook, R. J. (2001). Folate metabolism. In Homocysteine in Health and Disease (R. Carmel
and D. W. Jacobsen, eds.), pp. 113134. Cambridge University Press, Cambridge.
Corrales, F. J., Perez-Mato, I., Sanchez del Pino, M. M., Ruiz, F., Castro, C., Garcia-
Trevijano, E. R., Latasa, U., Martinez-Chantar, M. L., Martinez-Cruz, A., Avila, M. A.,
and Mato, J. M. (2002). Regulation of mammalian liver methionine adenosyltransferase.
J. Nutr. 132, 2377S2381S.
Curtin, K., Bigler, J., Slattery, M. L., Caan, B., Potter, J. D., and Ulrich, C. M. (2004).
MTHFR C677T and A1298C polymorphisms: Diet, estrogen, and risk of colon cancer.
Cancer Epidemiol. Biomarkers. Prev. 13, 285292.
78 H. F. Nijhout et al.
Czeizel, A. E. (2004). The primary prevention of birth defects: Multivitamins or folic acid?
Int. J. Med. Sci. 1, 5061.
Di Pietro, E., Wang, X. L., and MacKenzie, R. E. (2004). The expression of mitochondrial
methylenetetrahydrofolate dehydrogenase-cyclohydrolase supports a role in rapid cell
growth. Biochim. Biophys. Acta 1674, 7884.
Edelstein-Keshet, L. (1988). Mathematical Models in Biology. RandomHouse, New York.
Elowitz, M. B., Levine, A. J., Siggia, E. D., and Swain, P. S. (2002). Stochastic gene
expression in a single cell. Science 297, 11831186.
Fell, D. A. (1992). Metabolic control analysis: A survey of its theoretical and experimental
background. Biochem. J. 286, 313330.
Finkelstein, J. (1990). Methionine metabolism in mammals. J. Nutr. Biochem. 1, 228237.
Finkelstein, J. (2001). Regulation of homocysteine metabolism. In Homocysteine in
Health and Disease (R. Carmel and D. W. Jacobsen, eds.), pp. 9299. Cambridge
Univ. Press, Cambridge.
Finkelstein, J. (2003). Methionine metabolism in liver disease. Am. J. Clin. Nutr. 77,
10941095.
Finkelstein, J., and Martin, J. (1984). Methionine metabolism in mammals: Distribution of
homocysteine between competing pathways. J. Biol. Chem. 259, 95089513.
Finkelstein, J., and Martin, J. (1986). Methionine metabolism in mammals: Adaptation to
methionine excess. J. Biol. Chem. 261, 15821587.
Finkelstein, J. D., Harris, B. J., and Kyle, W. E. (1972). Methionine metabolism in
mammals: Kinetic study of betaine:Homocysteine methyltransferase. Arch. Biochem.
Biophys. 153, 320324.
Finkelstein, J. D., Kyle, W. E., and Harris, B. J. (1974). Methionine metabolism in
mammals: Regulatory effects of S-adenosylmethionine. Arch. Biochem. Biophys. 165,
774779.
Finkelstein, J. D., Harris, B. J., and Martin, J. J. (1982). Methionine metabolism in mammals:
Concentrations of metabolites in rat tissues. Kinetic study of betaine:homocysteine
methyltransferase. J. Nutr. 112, 10111018.
Frosst, P., Blom, H. J., Milos, R., Goyette, P., Sheppard, C. A., Matthews, R. G.,
Boers, G. J. H., Den Heijer, M., Kluijtmans, L. A. J., Van den Heuvel, L. P., and
Rozen, R. (1995). A candidate genetic risk factor for vascular diseaseA common
mutation in methylenetetrahydrofolate reductase. Nat. Genet. 10, 111113.
Fukushima, M., Morita, M., Ikeda, K., and Nagayama, S. (2003). Population study of
expression of thymidylate synthase and dihydropyrimidine dehydrogenase in patients
with solid tumors. Int. J. Mol. Med. 12, 839844.
Gregory, J. F., and Quinlivan, E. P. (2002). In vivo kinetics of folate metabolism. Annu. Rev.
Nutr. 22, 199220.
Gregory, J. F., Williamson, J., Liao, J. -F., Bailey, L. B., and Toth, J. P. (1998). Kinetic
model of folate metabolism in nonpregnant women consuming [
2
H
2
] folic acid: Isotopic
labeling of urinary folate and the catabolite para-acetamidobenzoylglutamate indicates
slow, intake-dependent, turnover of folate pools. J. Nutr. 128, 18961906.
Hartl, D. L., Dijkhuizen, D. E., and Dean, A. M. (1985). Limits of adaptation: The evolution
of selective neutrality. Genetics 111, 655674.
Harvey, R. J., and Dev, I. K. (1975). Regulation in the folate pathway of Escherichia coli. Adv.
Enzyme. Regul. 13, 99124.
Herbig, K., Chiang, E.-P., Lee, L.-R., Hills, J., Shane, B., and Stover, P. J. (2002).
Cytoplasmic serine hydroxymethyltransferase mediates competition between folate-
dependent deoxyribonucleotide and S-adenosylmethionine biosynthesis. J. Biol. Chem.
277, 3838138389.
International HapMap Consortium (2007). A second generation human haplotype map of
over 3.1 million SNPs. Nature 449, 851862.
Mathematical Models of Folate-Mediated One-Carbon Metabolism 79
Jackson, R. C. (1980). Kinetic simulation of anticancer drug interactions. Int. J. Biomed.
Comput 11, 197224.
Jackson, R. C. (1984). A kinetic model of regulation of the deoxyribonucleoside triphos-
phate pool composition. Pharmacol. Ther. 24, 279301.
Jackson, R. C. (1986). Kinetic simulation of anticancer drug effects on metabolic pathway
fluxesTwo case studies. Bull. Math. Biol. 48, 337351.
Jackson, R. C. (1993). The kinetic properties of switch antimetabolites. J. Nat. Cancer Inst.
85, 539545.
Jackson, R. C. (1995). Toxicity prediction from metabolic pathway modeling. Toxicology
102, 197205.
Jackson, R. C., and Harrap, K. R. (1973). Studies with a mathematical model of folate
metabolism. Arch. Biochem. Biophys. 158, 827841.
Jackson, R. C., and Harrap, K. R. (1979). Computer models of anti-cancer drug interaction.
Pharmacol. Ther. 4, 245280.
Janosik, M., Kery, V., Gaustadnes, M., Maclean, K. N., and Kraus, J. P. (2001). Regulation
of human cystathionine beta-synthase by S-adenosyl-L-methionine: Evidence for two
catalytically active conformations involving an autoinhibitory domain in the C-terminal
region. Biochemistry 40, 1062510633.
Jencks, D. A., and Matthews, R. G. (1987). Allosteric inhibition of methylenetetrahydro-
folate reductase by adenosylmethionine. Effects of adenosylmethionine and NADPH on
the equilibrium between active and inactive forms of the enzyme and on the kinetics of
approach to equilibrium. J. Biol. Chem. 262, 24852493.
Kastanos, E. K., Woldman, Y. Y., and Appling, D. R. (1997). Role of mitochondrial and
cytoplasmic serine hydroxymethyltransferase isozymes in de novo purine synthesis in
Saccharomyces cerevisiae. Biochemistry 36, 1495614964.
Kluijtmans, L. A. J., Boers, G. H. J., Stevens, E. M. B., Renier, W. O., Kraus, J. P.,
Trijbels, F. J. M., Van den Heuvel, L. P. W. J., and Blom, H. J. (1996). Defective
cystathionine beta-synthase regulation by S-adenosylmethionine in a partially pyridoxine
responsive homocystinuria patient. J. Clin. Invest. 98, 285289.
Leduc, D., Escartin, F., Nijhout, H. F., Reed, M. C., Liebl, U., Skouloubris, S., and
Myllykallio, H. (2007). Flavin-dependent thymidylate synthase ThyX activity: Implica-
tions for the folate cycle in bacteria. J. Bacteriol. 189, 85378545.
MacFarlane, A. J., Perry, C., Liu, X., Allen, S., and Stover, P. J. (2005). Cytoplasmic serine
hydroxymethyltransferase regulates the homocysteine remethylation pathway. Haematol.
Rep. 1, 28.
Martinov, M., Vitvitsky, V., Mosharov, E., Banerjee, R., and Ataullakhanov, F. (2000).
A substrate switch: A new mode of regulation in the methionine metabolic pathway.
J. Theor. Biol. 204, 521532.
McVean, G., Spencer, C. C. A., and Chaix, R. (2005). Perspectives on human genetic
variation from the HapMap project. PLOS Genet. 1, e54.
Mejia, N. R., and MacKenzie, R. E. (1986). NAD-dependent methylenetetrahydrofolate
dehydrogenase-methylenetatrahydrofolate cyclohydrolase in transformed cells is a mito-
chondrial enzyme. Biochem. Biophys. Res. Commun. 155, 16.
Morris, M. C., Evans, D. A., Bienias, J. L., Tangney, C. C., Herbert, L. E., Scherr, P. A., and
Schneider, J. A. (2005). Dietary folate and vitamin B12 intake and cognitive decline
among community-dwelling older persons. Arch. Neurol. 62, 641645.
Morrison, P. F., and Allegra, C. J. (1989). Folate cycle kinetics in human breast cancer cells.
J. Biol. Chem. 264, 1055210566.
Murray, J. D. (1989). Mathematical Biology. Springer-Verlag, Berlin.
Nijhout, H. F., Reed, M. C., Budu, P., and Ulrich, C. M. (2004). A mathematical model of
the folate cycle. J. Biol. Chem. 279, 5500855016.
80 H. F. Nijhout et al.
Nijhout, H. F., Reed, M. C., Anderson, D. F., Mattingly, J. C., James, S. J., and
Ulrich, C. M. (2006). Long-range allosteric interactions between the folate and methio-
nine cycles stabilize DNA methylation reaction rate. Epigenetics 1, 8187.
Nijhout, H. F., Reed, M. C., Lam, S.-L., Shane, B., Gregory, J. F., and Ulrich, C. M.
(2007a). In silico experimentation with a model of hepatic mitochondrial folate metabo-
lism. Theor. Biol. Med. Model. 3, 40ff.
Nijhout, H. F., Reed, M. C., and Ulrich, C. M. (2007b). A day in the life of cell metabolism.
Biol. Theor. 2, 124127.
Obama, K., Kanai, M., Kawai, Y., Fukushima, M., and Takabayashi, A. (2002). Role of
retinoblastoma protein and E2F-1 transcription factor in the acquisition of 5-fluorouracil
resistance by colon cancer cells. Int. J. Oncol. 21, 309314.
Ou, X., Yang, H., Ramani, K., Ara, A. I., Chen, H., Mato, J. M., and Lu, S. C. (2007).
Inhibition of human betainehomocysteine methyltransferase expression by
S-adenosylmethionine and methylthioadenosine. Biochem. J. 401, 8796.
Peri, K. G., Belanger, C., and MacKenzie, R. E. (1989). Nucleotide sequence of the human
NAD-dependent methylenetetrahydrofolate dehydrogenase-cyclohydrolase. Nucleic
Acids Res. 17, 8853.
Pogribna, M., Melnyk, S., Pogrybny, I., Chango, A., Yi, P., and James, J. (2001). Homo-
cysteine metabolism in children with down syndrome: In vitro modulation. Am. J. Hum.
Genet. 69, 8895.
Prudova, A., Martinov, M. V., Vitvitsky, V. M., Ataullakhanov, F. I., and Banerjee, R.
(2005). Analysis of pathological defects in methionine metabolism using a simple mathe-
matical model. Biochim. Biophys. Acta 1741, 331338.
Reed, M. C., Nijhout, H. F., Sparks, R., and Ulrich, C. M. (2004). A mathematical model
of the methionine cycle. J. Theor. Biol. 226, 3343.
Reed, M. C., Nijhout, H. F., Neuhouser, M. L., Gregory, J. F., Shane, B., James, S. J.,
Boynton, A., and Ulrich, C. M. (2006). A mathematical model gives insights into
nutritional and genetic aspects of folate-mediated one-carbon metabolism. J. Nutr. 136,
26532661.
Reed, M. C., Thomas, R. L., Pavisic, J., James, S. J., Ulrich, C. M., and Nijhout, H. F.
(2008). A mathematical model of glutathione metabolism. Theoret. Biol. Math. Model.
(in press).
Rockman, M. V., and Wray, G. A. (2002). Abundant raw material for cis-regulatory
evolution in humans. Mol. Biol. Evol. 19, 19912004.
Rodriguez-Caso, C., Montanez, R., Cascante, M., Sanchez-Jimenez, F., and Medina, M. A.
(2006). Mathematical modeling of polyamine metabolism in mammals. J. Biol. Chem.
281, 2179921812.
Rosenblatt, D. (2001). Inborn errors of folate and cobalamin metabolism. In Homocysteine
in Health and Disease (R. Carmel and D. W. Jacobsen, eds.), pp. 244258. Cambridge
Univ. Press, Cambridge.
Segel, I. H. (1975). Enzyme Kinetics. John Wiley & Sons, New York.
Seither, R. L., Trent, D. F., Mikulecky, D. C., Rape, T. J., and Goldman, I. D. (1989).
Folate pool interconversions and inhibition of biosynthetic processes after exposure of
L1210 leukemia cells to antifolatesExperimental and network thermodynamic analyses
of the role of dihydrofolate polyglutamylates in antifolate action in cell. J. Biol. Chem.
264, 1701617023.
Shane, B. (1995). Folate chemistry and metabolism. In Folate in Health and Disease
(L. B. Bailey, ed.), pp. 122. Marcel Dekker, New York.
Sigal, A., Milo, R., Cohen, A., Naama, G. -Z., Klein, Y., Liron, Y., Rosenfeld, N.,
Danon, T., Perzov, N., and Alon, U. (2006). Variability and memory of protein levels
in human cells. Nature 444, 643646.
Mathematical Models of Folate-Mediated One-Carbon Metabolism 81
Silberberg, J., and Dudman, N. (2001). Methionine loading. In Homocysteine in Health
and Disease (R. Carmel and D. W. Jacobsen, eds.), pp. 212219. Cambridge University
Press, Cambridge.
Slansky, J. E., Li, Y., Kaelin, W. G., and Farnham, P. J. (1993). A protein synthesis-
dependent increase in E2F1 messenger-RNA correlates with growth regulation of the
dihydrofolate reductase promoter. Mol. Cell. Biol. 13, 16101618.
Smith, G. K., Banks, S. D., Monaco, T., Rigual, R., Duch, D. S., Mullin, R. J., and
Huber, B. E. (1990). Activity of an NAD-dependent 5,10-methylenetatrahydrofolate
dehydrogenase in normal tissue, neoplastic cells, and oncogene-transformed cells. Arch.
Biochem. Biophys. 283, 367371.
Sunder-Plassman, G., Foedinger, M., Buchmayer, H., Papagiannopoulos, M., Wojcik, J.,
Kletzmayr, J., Enzenberger, B., Janat, A. O., Winkelmayer, W. C., Paul, G.,
Auinger, M., Barnas, U., et al. (2000). Effect of high dose folic acid therapy on hyperho-
mocysteinemia in hemodialysis patients: Results of the Vienna Multicenter Study. J. Am.
Soc. Nephrol. 11, 11061116.
Troen, A. M., Mitchell, B., Sorensen, B., Wener, M. H., Johnston, A., Wood, B.,
Selhub, J., McTiernan, A., Yasui, Y., Potter, J. D., and Ulrich, C. M. (2006). Unmetab-
olized folic acid in plasma is associated with reduced natural killer cell cytotoxicity among
postmenpausal women. J. Nutr. 136, 189194.
Ulrich, C. M., Neuhouser, M., Lui, A. Y., Boynton, A., Gregory, J. F. III, Shane, B.,
James, S. J., Reed, M. C., and Nijhout, H. F. (2008). Mathematical modeling of
folate metabolism: Predicted effects of genetic polymorphisms on mechanisms and
biomarkers relevant to carcinogenesis. Cancer Epidemiol. Biomark. Prev. (in press).
Vorontzov, I. N., Greshilov, M. M., Belousova, A. K., and Gerasimova, G. K. (1980).
Mathematical description and investigation of the principles of functioning in the folic
acid cycle. Biokhimiya 45, 8397.
Wade, M., Blake, M. C., Jambou, R. C., Helin, K., Harlow, E., and Azizkhan, J. C. (1995).
An inverted repeat motif stabilizes binding of E2F and enhances transcription of the
dihydrofolate reductase gene. J. Biol. Chem. 270, 97839791.
Wagner, A. (2005). Robustness and Evolvability in Living Systems. Princeton Univ.
Press, Princeton NJ.
Wagner, C. (1995). Biochemical role of folate in cell metabolism. In Folate in Health and
Disease (L. B. Bailey, ed.), pp. 2342. Marcel Dekker, New York.
Wagner, C., Briggs, W. T., and Cook, R. J. (1985). Inhibition of glycine
N-methyltransferase activity by folate derivatives: Implications for regulation of methyl
group metabolism. Biochem. Biophys. Res. Comm. 27, 746752.
Werkheiser, W. C., Grindey, G. B., Moran, R. G., and Nichol, C. A. (1973). Mathematical
simulation of the interaction of drugs that inhibit deoxyribonucleic acid biosynthesis.
Molec. Pharmacol. 9, 320329.
Woeller, C. F., Anderson, D. D., Doletha, Szebenyi, M. E., and Stover, P. J. (2007).
Evidence for small ubiquitin-like modifier-dependent nuclear import of the thymidylate
biosynthesis pathway. J. Biol. Chem. 282, 1762317631.
Yamada, K., Chen, Z., Rozen, R., and Matthews, R. G. (2001). Effects of common
polymorphisms on the properties of recombinant human methylenetetrahydrofolate
reductase. Proc. Nat. Acad. Sci. USA 98, 1485314858.
Yan, H., Yuan, W., Velculescu, V. E., Vogelstein, B., and Kinzler, K. W. (2002). Allelic
variation in human gene expression. Science 297, 1143.
Yeo, E. J., and Wagner, C. (1992). Purification and properties of pancreatic glycine
N-methyltransferase. J. Biol. Chem. 267, 2466924674.
82 H. F. Nijhout et al.
C H A P T E R T H R E E
Folate Deprivation, the Methionine
Cycle, and Alzheimers Disease
Flaubert Tchantchou* and Thomas B. Shea
Contents
I. Introduction 84
II. Folate Metabolism, the Transmethylation Pathway, and AD 85
III. The Transsulfuration PathwayHomocysteine Elimination and
Glutathione Metabolism 90
References 94
Abstract
Folate deficiency is associated with increase in homocysteine levels. Abnormal
plasma levels of that neurotoxic nonproteinogenic amino acid is implicated in
many pathological conditions including cardiovascular diseases, neural tube
defects, and is now recognized as a risk factor in Alzheimers disease (AD)
dementia. Homocysteine elimination is regulated by two metabolic pathways,
namely, the transmethylation and the transsulfuration pathways. Its elimi-
nation via these two metabolic pathways is modulated by folate, a member
of the B-vitamin family. Folate provides, via its metabolic end product
5-methyltetrahydrofolate, a methyl group that is used to reconvert homocyste-
ine back to methionine through the transmethylation pathway. The efficiency of
folate metabolism has an impact on the availability of S-adenosylmethionine,
a compound that is known to activate homocysteine flux through the transsul-
furation pathway and is necessary for utilization of a downstream antioxidant
called glutathione under the catalysis of glutathione S-transferase enzyme.
In this review, we will explore the impact of folate deprivation on the regulation
of the methionine cycle and exhaustively describe different biochemical reac-
tions that are implicated in the regulation of homocysteine elimination and that
folate deficiency influences in AD neuropathology. 2008 Elsevier Inc.
Vitamins and Hormones, Volume 79
#
2008 Elsevier Inc.
ISSN 0083-6729, DOI: 10.1016/S0083-6729(08)00403-2 All rights reserved.
* University of Maryland, Baltimore, Maryland
{
Center for Cellular Neurobiology and Neurodegeneration Research, UMassLowell,
Lowell, Massachusetts 01854
83
I. Introduction
Folate is a member of the B-vitamin family and a carrier of one-carbon
fragments, which it transfers to various biochemical targets. Folate is impor-
tant for the functioning of the central nervous system (CNS) at all ages
(Bottiglieri et al., 1995; Reynolds, 2002). Its metabolism provides a methyl
group, via its metabolite 5-methyltetrahydrofolate, which is necessary for
the remethylation of the neurotoxic amino acid homocysteine back to
methionine, an essential amino acid that plays a key role in the generation
of methyl groups required for numerous biochemical reactions. Substantial
scientific evidence associates folate deficiency to Alzheimers disease (AD).
The deficiency of this B vitamin induces homocysteine accumulation. This
sulfur-containing nonproteinogenic amino acid transitionally exists at the
intersection of the transmethylation and the transsulfuration pathways,
which regulate its elimination (Selhub, 1999).
Hyperhomocysteinemia is associated with an increased risk of several
pathological conditions including vascular diseases and vascular dementia
and has been confirmed in patients with AD and mild cognitive impairment
(MCI) where they represent an independent risk factor (Seshadri et al.,
2002; Shea and Rogers, 2002a). HCY levels in several biological fluids
and tissues represent a predictive index for the incidence of AD and other
dementias. Substantial evidence has established a connection between HCY
metabolism and cognitive function. Abnormal levels of HCY have been
related to multiple cognitive dysfunctions including age-related memory
loss, vascular dementia, and AD (Malaguarnera et al., 2004; OSuilleabhain
et al., 2004; Sachdev et al., 2003). Deficiencies in folic acid are often
observed in the elderly population with a resultant increase in HCY.
They are proposed to be owing to an increasing prevalence of atrophic
gastritis type B, which occurs with a frequency of up to 50% in elderly
subjects (Wolters et al., 2004). The link between increase in homocysteine
levels and AD resulted from the growing recognition that cerebrovascular
disease may promote AD. This idea was taken from studies of HCY and
heart disease research and is being extended to cerebral disorders. This
correlation lays on the fact that plasma HCY maybe directly toxic to
vascular endothelial cells or induces their dysfunction, leading to the loss
of the bloodbrain-barrier function and altered production of nitric oxide.
In addition, HCY crossing the bloodbrain barrier or being released by cells
within the brain could act as a potent neurotoxin (Miller, 1999).
Such neurotoxic effects may be due to the direct interaction of HCY
with plasma membrane components or to the intracellular accumulation
of S-adenosylhomocysteine (SAH). This latter metabolite inhibits the
methylation of catechol substrates resulting in the generation of oxyradicals
and other chemically reactive products that are cytotoxic. Moreover,
84 Flaubert Tchantchou and Thomas B. Shea
homocysteine as sulfhydryl compound is an electron donor, which acts with
the transition metal ions, iron and copper, to generate hydrogen peroxide
(Kruman et al., 2002). HCY also has the ability to induce the storage of iron
from ferritin, and this could explain the increase in redox-active iron in AD
neurons and concomitant oxidative stress, which subsequently triggers
deposition of amyloid plaques in the AD brain (Ulrich et al., 2002).
Considering the deleterious effect of HCY accumulation in the brain, its
continuous elimination is necessary, and hence, the importance of the
methyl group provided by the folate (also known as vitamin B
9
) as it
provides 5-methyltetrahydrofolate, required for the reconversion of HCY
to methionine via the transmethylation pathway.
Another major consequence of folate deficiency is a decline in
S-adenosylmethionine (SAM; the major methyl donor). This decline in
SAM, which is endogenously generated from methionine, is responsible for
increased DNA breakage in mouse models (Kruman et al., 2000, 2002) and
the gradual hypomethylation of DNA accompanies aging and AD
(Morrison, 1996; Seshadri et al., 2002). The depletion of SAM can also
lead to overexpression of presenilin-1 (PS-1; Fuso et al., 2005; Scarpa et al.,
2003), which is associated to abnormal processing of the amyloid precursor
protein that results into the formation of the b-amyloid protein (Parihar and
Hemnani, 2004). Furthermore, this principal methyl donor mediates the
enzymatic reaction utilizing an endogenous antioxidant and downstream
metabolite of HCY metabolism via the transsulfuration pathway called
glutathione, under the catalysis of glutathione S-transferase enzyme
(Tchantchou et al., 2006b). Therefore, the utilization of glutathione by
glutathione S-transferase would promote HCY elimination via the trans-
sulfuration pathway. This clearly highlights the important role that folate
plays in the elimination of HCY via both the transmethylation and the
transsulfuration pathways.
II. Folate Metabolism, the Transmethylation
Pathway, and AD
The transmethylation pathway is derived from the conjunction of two
biochemical pathways, namely, the folate and the methionine metabolic
pathways. The transmethylation pathway consists of transferring a methyl
group (CH
3
) to HCY by the end product of folate metabolism,
5-methyltetrahydrofolate, or betaine to form methionine. Folate is a mem-
ber of the B-vitamin family and a carrier of one-carbon fragments, which it
transfers to various biochemical targets (Chen et al., 2004). Its metabolism
starts with its deconjugation in the cells of the intestinal wall to the mono-
glutamate form. This form is further reduced to dihydrofolate and then to
Folate Deprivation, the Methionine Cycle, and Alzheimers Disease 85
tetrahydrofolate (THF) via the catalytic actions of folate and dihydrofolate
reductase enzymes, respectively (Fig. 3.1; reactions 1 and 2, respectively).
Both of these enzymes require NAPDH as a cofactor. The resulting THF
receives a hydroxymethyl group from the combination of a serine molecule
witha pyridoxal-5
0
-phosphate toform5,10-methylene tetrahydrofolate (5,10-
methyleneTHF) and glycine in a reaction catalyzed by the enzyme
serine hydroxymethyltransferase (Fig. 3.1, reaction 3). 5,10-MethyleneTHF
NADPH + H
+
NADP
+
Diet intake
(Circulating folate)
Folate
Cystathionine
Cysteine
5,10-methylene THF
SAM
SAH
5-MeTHF
PPi + Pi
ATP
CH
3
-
CH
3
-Ac
Adenosyl
a-KB
GSH
2H
2
O
GR; Rx-13
GS; Rx-11
GPx; Rx-12
CBS/Vit B
6
; Rx-9
MS/Vit B12; Rx-5
MTHFR; Rx-4
MAT; Rx-6
SAHH; Rx-8
SAMT; Rx-7
GSH-E-CDNB
GsT(E); Rx-14
CDNB
E + P
Serine
NADPH + H
+
NADP
+
FR; Rx-1
Dihydrofolate
THF
NADPH + H
+
NADP
+
DR; Rx-2
Serine
Glycine
Homocysteine
Cg L;
Rx-10
H
2
O
2
Choline
Betaine
Acetylcholine
Acetyl-CoA
Methionine
ChaT, Rx15
GS-SG
HMT; Rx-3
Figure 3.1 Pathway regulating homocysteine elimination and glutathione metabolism
in the brain.
86 Flaubert Tchantchou and Thomas B. Shea
is of central importance in several biological events. It is the precursor of
the metabolically active 5-methyltetrahydrofolate, which is involved in
HCY metabolism, and methylene tetrahydrofolate, which is involved in
purine synthesis (Brotto and Yang, 2000; Raber et al., 1998; Selhub, 1999).
With relevance to the transmethylation pathway, 5,10-methyleneTHF
produces 5-MTHF in a reaction catalyzed by methylene tetrahydrofolate
reductase (MTHFR) (Fig. 3.1, reaction 4). Variations in the gene encoding
for the MTHFR enzyme can decrease folate metabolism and subsequently
result in an increase in HCY levels. Patients with congenital MTHFR
deficiency have reduced levels of several important biological metabolites
such as methionine and SAM in the cerebrospinal fluid (CSF) and show
demyelination in the brain which might be due to decreased methylation.
MTHFR polymorphisms that exhibit decreased activity are present in as
much as 20% of some populations (Brotto and Yang, 2000). Diminished
activity of this enzyme also reduces the production of SAM (required for
DNA methylation). MTHFR deficiency and the presence of ApoE4 may
represent synergistic risk factors for AD. This is a plausible extension of the
known association of diminished folate metabolism with AD, in that indi-
viduals who are homozygote carriers of MTHFR polymorphic gene are at
particular risk for other folate-related neural defects when plasma folate is at the
low end of the normal range (Shields, 1999). Moreover, individuals
affected with AD who are also homozygote carriers of MTHFR poly-
morphisms show elevated HCY levels compared to nonhomozygous
ADpatients despite the presence of equal levels inother considerable biological
compounds such as folate in both groups. It has also been demonstrated that
mice heterozygous for the MTHFR defect are vulnera-
ble to hyperhomocysteinemia when fed with lowfolate diets and have altered
tissue methylation capacity and impaired endothelial function in
cerebral microvessels (Devlin et al., 2004). APOE
/
, APOE
/
, and
APOE
/
mice exposed to oxidative stress inducing diet deficient in folate
show increased transcription and activity levels of MTHFR when com-
pared to APOE
/
mice maintained on folate supplemented diet. This is
indicative of the need for folate, whose metabolic product 5-MTHF is the
principal methyl donor for remethylation of HCY back to methionine in
the CNS in a reaction catalyzed by methionine synthase (MS) (also known
as 5-methyltetrahydrofolatehomocysteine S-methyltransferase; Fig. 3.1, reac-
tion 5), as betaine, the alternative methyl donor, is absent in the brain. Previous
studies showed that folate supplementation with or without vitamins B
6
or B
12
to individuals with hyperhomocysteinemia lower their homocysteine levels
(Anonymous, 1998), and recently, a 2-year term double-blind placebo-
controlled randomized clinical trial showed no association between homocys-
teine lowering and cognitive performance (McMahon et al., 2006). The
methodological quality of that study is questionable in that the investigators
recruited for their trial, healthy elderly people who were showing no signs of
Folate Deprivation, the Methionine Cycle, and Alzheimers Disease 87
cognitive dysfunction and had no means of determining whether they were
prone to developing cognitive impairment during the trial. The erroneous
sampling of the targeted population rendered the results of their findings
opened to serious critical comments. During homocysteine reconversion to
methionine, THF can be regenerated as the second product of the reaction.
This reaction is a precursor to the regeneration of 5-MTHF via 5,10-methy-
leneTHF (Bronstrup et al., 1998). Folate deficiency mediates neurotoxicity in
part by increasing levels of HCY. This nonproteinogenic amino acid over-
stimulates N-methyl-D-aspartate (NMDA) receptors, potentiates glutamate
accumulation and amyloid-b aggregation and neurotoxicity, and induces
DNA breakage and lipid peroxidation (Al-Gazali, 2001; Ho, 2001). Mouse
models of AD and Parkinsons disease as well as wild-type mice subjected to
folate deficiency show elevated HCY and place neurons at the risk of
degeneration and endothelial damage. The mechanism whereby homo-
cysteine leads to endothelial cell damage has been found to be via its
auto-oxidation to homocysteine and H
2
O
2
(Loscalzo, 1996). Electron
microscopic studies of the brains of folate-deprived rats revealed that
hyperhomocysteinemia induced by folate deprivation was accompanied by
ultrastructural degenerative changes in the cerebral microvasculature,
including endothelial and pericytic degeneration, mitochondrial destruc-
tion, and cytoplasmic dissolution (Kim et al., 2002). These alterations were
similar to the previously reported microvascular degenerative features typi-
cally found in cerebral diseases such as AD, Parkinsons disease, and aging
processes (Farkas et al., 2000; Ureno et al., 2001). Oxidative damage due to
folate deficiency is potentiated by lack of apolipoprotein E gene and iron
supplementation as pro-oxidant. The extent of the damage to brain cells can
be determined by the measure of thiobarbituric acid reactive species
(TBARs) levels, which is an index of oxidative damage induced by lipid
peroxidation (Ho et al., 2003; Mattson and Haberman, 2003; Shea et al.,
2001). The expression status of MS is easily altered under oxidative stress.
The transcription levels and/or the activity of the enzyme it encodes for is
significantly decreased under different oxidizing conditions. Vitamin B
12
is an indispensable cofactor in the transmethylation reaction in the brain. This
reaction is of great importance inthe regulation of serumHCYlevels andis the
only reaction in the body in which folate and vitamin B
12
are co-participants
(Mattson et al., 2002; Miller and Kelly, 1997). A decrease in MS transcription
andactivityobservedinnormal mice under oxidizingconditions canbe viewed
as a natural downregulating compensatory process, which attempts to avoid
further regeneration of methionine from HCY, which would infinitely be
demethylated to reform HCY. Consistent with this line of thought, other
investigations have shown that the increased HCY flux through the transsul-
furation pathway could result from an increase in the levels of methionine
adenosyltransferase and/or cystathionine-b-synthase or a decrease in methio-
nine synthase activity (Mosharov et al., 2000; Tchantchou et al., 2006a). The
88 Flaubert Tchantchou and Thomas B. Shea
methyl group transferred from 5-MTHF to HCY to form methionine con-
tributes to the formation of SAM in a one-step reaction in which an ATP
molecule is involved (Fig. 3.1, reaction 6) and that represents the end point of
the transmethylationpathway. Inthe reactionmechanismof the remethylation
of HCYtomethionine, a methyl groupis transferredtothe MScofactor, cob(I)
alamin, whichis thenactivatedbyforming methylcobalamin. The activationof
cob(I)alamin to form methylcobalamin requires the methyl donor SAM. But
once the first molecule of methylcobalamin is formed and used for the recon-
versionof HCYto methionine, subsequent molecules couldbe regenerated by
using 5-MTHF as methyl donor to serve the same purpose. The absence of the
alternative betaine remethylation pathway in the CNS greatly reduces the
methylation capacity. Therefore, folate deprivation would inhibit transmethy-
lation reactions by reducing SAM and further potentiates HCY accumulation
in the CNS. Considering that the action of vitamin B
12
plays a role in HCY
metabolism that is similar to that of folate (Mattson and Shea, 2003), and
because folate and vitamin B
12
deficiencies retard methionine regeneration,
SAM levels are also reduced as a consequence of lack of folate or deficiency in
vitamin B
12
action (Selhub and Miller, 1992). Impaired methylation has been
implicated in many neurological and psychological disorders, including
dementia, depression, and psychosis. Decreased intracellular methylation reac-
tions canalsoresult inanincrease of SAH. This line of reasoningis supportedby
the demonstration that HCY induces DNA breakage and resultant apoptosis,
and that co-treatment with SAMprevented homocysteine-induced apoptosis.
The participation of SAMin methylation reactions results in the production of
SAHunder the catalysis of S-adenosylmethionine methyltransferase (Cantoni,
1986) (Fig. 3.1, reaction7). The hydrolysis of SAHcatalyzedby SAHhydrolase
produces HCYand adenosine (Fig. 3.1, reaction 8). The compensatory down-
regulation of the reconversion of homocysteine to methionine andthe absence
of the betaine metabolic pathway inthe brain further excludes the possibility of
efficiently regenerating methionine from homocysteine. Choline, a precursor
of betaine in other organs, could therefore condense with acetyl-coA to
optimally produce the neurotransmitter acetylcholine inthe brain, ina reaction
catalyzed by choline acetyltransferase (ChaT) (Fig. 3.1, reaction 15) (Fisher
et al., 2002). Findings of a recent study demonstrates that dietary folate defi-
ciency in middle age adult mice decreased levels of acetylcholine, which were
restored by dietary supplementation with SAM even in the absence of folate.
Adult mice heterozygously lacking 5
0
,10
0
-methylene tetrahydrofolate reduc-
tase (which are impaired in folate usage and have reduced SAM levels), or
homozygously lacking apolipoprotein E (which have reduced SAM due to
oxidativestress) andagednormal miceeachdisplayreducedacetylcholine levels
as compared to normal adult mice. Dietary folate deficiency further reduced
acetylcholine and induced cognitive impairment in each of these mice, while
supplementation with SAMin the absence of dietary folate restored acetylcho-
line levels andcognitive performancetolevels observedinthe presenceof folate
Folate Deprivation, the Methionine Cycle, and Alzheimers Disease 89
(Chan et al., 2006). That observation underscores the importance of folate and
the principal methyl donor SAM in the regulation of acetylcholine levels,
which represents an important the rapeutic approach for age-related neurode-
generation such as AD.
III. TheTranssulfurationPathwayHomocysteine
Elimination and Glutathione Metabolism
The transsulfuration pathway is another path for HCY elimination via
which about 60% of HCY is believed to be metabolized (Storch et al.,
1988). It comprises several reaction sequences which start with the forma-
tion of cystathionine followed by those of cysteine and glutathione.
Cystathionine formation is derived from the condensation of L-serine
with homocysteine, in a reaction catalyzed by the heme and vitamin B
6
-
dependent cystathionine-b-synthase (CBS). Cystathionine is subsequently
cleaved to yield cysteine and 2-ketobutyrate in a reaction catalyzed by
cystathionine g-lyase (Mosharov et al., 2000). But, cystathionine g-lyase is
absent or has a limited presence in the brain therefore, levels of cystathio-
nine found in brain tissues is putatively due to the action of CBS since it is
considered the rate-limiting enzyme in HCY transsulfuration. CBS genetic
knockout mice model first developed by Watanabe et al. (1995) exhibit
increased HCY levels. Among them, CBS homozygote knockout mice
developed significantly elevated levels of plasma total HCY compared
to heterozygote knockout which showed mildly elevated concentration
of total plasma HCY, which are closely similar to those in humans
with heterozygous CBS deficiency. Those findings suggest that partial
impairment in homocysteine transsulfuration produces similar effects on
HCY metabolism in humans and mice. A recent study aiming at evalu-
ating the in vivo effect of high serum homocysteine concentration on
amyloid-b-peptide (Ab) levels in the brain and in relation to AD neuropa-
thology using a mice model carrying the well-established amyloidosis
mutant genes APP/PS1 (Holcomb et al., 1998) and heterozygous for a
cystathionine-b-synthase mutation (APP/PS1/CBS); therefore, resulting
in deficient CBS activity and high homocysteine levels. The mouse
model showed significant elevations of serum homocysteine levels com-
pared to the double transgenic APP/PS1 model of amyloidosis. Results
showed that female (but not male) APP/PS1/CBS mice exhibited signifi-
cant elevations of Ab40 and Ab42 levels in the brain. Correlations between
homocysteine levels in serum and brain Ab levels were statistically signifi-
cant (Pacheco-Quinto et al., 2006). Deficiency in CBS leading to
90 Flaubert Tchantchou and Thomas B. Shea
homocysteinuria results in multiple organs/systems damage, severe vascular
disease, and mental retardation (Mudd et al., 1995). Human hepatoma cell
line demonstrates increase synthesis of cystathionine under oxidative stress
conditions. The increase synthesis of cystathionine is followed by an
increase in HCY flux through the transsulfuration pathway (Mosharov
et al., 2000). Increased transcription levels of CBS gene are present in
apolipoprotein E homozygote and heterozygote knockout mice in a
gene dosage manner. This increase is potentiated by folate deprivation
(Tchantchou et al., 2006a). The immediate consequence of this increase
in cystathionine levels is that it could lead to an increase in the concentra-
tions of the downstream metabolites, cysteine and glutathione (GSH). The
transsulfuration reaction thus provides a direct link between homocysteine
and glutathione, the major endogenous redox buffer in mammalian cells.
It is therefore not surprising that a number of enzymes at this metabolic
nexus display sensitivity to redox changes (Mosharov et al., 2000). Thus,
changes in the levels of expression or functional activity of CBS can
affect levels of HCY (Mattson and Shea, 2003). Such a regulatory switch
could be rationalized as representing a self-correcting response to depleted
glutathione levels in cells faced with an oxidative challenge (Loehrer et al.,
1996). This highlights the overwhelming importance of the transsulfuration
pathway, which under oxidative stress has a dual beneficial impact
since its upregulation would, in addition to accelerating homocysteine
elimination, also contribute to the increase synthesis of the antioxidant
glutathione in many cell and tissue types. In so doing, the transsulfuration
pathway contributes at least indirectly in preventing or quenching oxidative
damage to the brain and other organs. GSH systematically called g-gluta-
mylcysteinylglycine is a ubiquitous tripeptide, formed from the amino acids
glutamate, glycine, and cysteine by two ATP-dependent enzymatic reac-
tions (Schulz et al., 2000). GSH is a major intracellular antioxidant and
its antioxidant activity depends upon the thiol group within the molecules.
GSH is crucial in the free radical scavenging of singlet oxygen and the
OH radical (Larsson et al., 1983). GSH plays this crucial role in detoxifying
peroxides and/or electrophilic toxins such as 1-chloro-2,4-dinitrobenzene
(CDNB) used as substrates in reactions catalyzed by GSH peroxidase
(Fig. 3.1, reaction 12) and glutathione S-transferase (Fig. 3.1, reaction 14),
respectively. Intracellular GSH is maintained in its thiol form by glutathione
disulfide (GSSG) reductase, which requires NADPH molecule (Fig. 3.1,
reaction 13). The availability of cysteine is critical for the synthesis of
GSH in most cells (Ceballos-Picot et al., 1996). Glutathione levels and acti-
vity of glutathione synthase (GS) are increased under oxidative stress
conditions induced by dietary deficiency (folate and vitamin E). This increase
in glutathione levels is substantiated by apoliprotein E deficiency (Gilgun-
Folate Deprivation, the Methionine Cycle, and Alzheimers Disease 91
Sherki et al., 2001; Shea et al., 2002). Deficiency in this gene was shown to
promote the increase of oxidative stress (Ramassamy et al., 1999). Moreover,
experimental elevations in glutathione in AD brain were capable of reducing
oxidative damage and therefore represented an attempt to compensate for
increased ROS induced by dietary and genetic deficiency (Huang et al., 2000).
Furthermore, this increase in GSH levels in the nervous system is triggered by
the upregulation of the transcription and activity profile of glutathione
synthase. Glutathione synthase gene and activity displayed differential
compensatory responses to dietary folate and apolipoprotein E deficiency.
A significant increase in transcription of GS is observed only in APOE
/
mice and only when they were maintained on folate-deficient diet, suggesting
that the combined impact of diet-induced and genetically induced oxidative
stress is required to induce an increase in transcription. The magnitude of this
combined impact is reflected by a synergistic increase in thiobarbituric acid-
reactive substance levels in brain tissue of APOE
/
mice maintained under
folate deficiency. Maintenance of normal mice ondietary folate deficiency also
induces a significant increase in the combined impact of the absence of APOE
and the dietary folate deficiency results in a dramatic increase in TBARs in
brain tissue (Shea and Rogers, 2002b), and further indicates a synergistic
deleterious impact of these dietary folate and genetic deficiencies. Therefore,
the increase in GS transcription and activity in APOE
/
mice subjected to
oxidative stress inducing diet correlate with the synergistic increase in TBARs.
But, these increases in both activity and transcription of GS in the brain of
APOE
/
mice maintained on folate-deficient diet are unable to compensate
fully for the synergistic increase in oxidative damage. This observation under-
scores the extent of oxidative damage that diet-induced and genetically
induced oxidative stress could cause to brain tissue (Tchantchou et al.,
2004a). These findings highlight the fact that distinct compensatory responses
in an antioxidant-generating enzyme can be invoked depending on the nature
and extent of oxidative stress. The combined efficacy of these responses was
reflected by steady-state levels of glutathione, in that both diet-induced and
genetically induced oxidative stress individually elevated glutathione levels,
whereas the combined impact of both induced an apparent additive increase
(Shea and Rogers, 2002b; Tchantchou et al., 2004a). The cumulative increase
of GSHlevels in brain tissues under oxidative damage is suggestive of a possible
alteration of the activity of enzymes that help use glutathione to quench
reactive oxygen species and toxins that induce oxidative damage to the
brain. A recent clinical trial with a small number of cognitively impaired
patients demonstrated that the therapeutic combination of N-acetylcysteine
and B-vitamin supplements (folate and vitamin B
12
) improved cognitive
status of these hyperhomocysteinemic patients (McCaddon, 2006). In that
combinatorial therapy, which has homocysteine levels lowering properties,
NAC will increase the flow of homocysteine through the transsulfuration
92 Flaubert Tchantchou and Thomas B. Shea
pathway leading to an increase in GSHformation while folate and vitamin B
12
will regulate homocysteine remethylation via the transmethylation pathway.
The use of N-acetylcysteine in this combinatorial therapy correlates with
previous findings demonstrating that the administration of N-acetylcysteine
to ApoE-deficient mice deprived of folate alleviated oxidative damage and
cognitive decline, and their restored glutathione synthase and GSH levels to
those of normal mice maintained in the presence of folate (Tchantchou et al.,
2005). The activity and the transcription profile of GSH peroxidase which
catalyzes the reaction in which GSH is used to eliminate hydrogen peroxide
and results in the formation of the oxidized form of glutathione (GSSG) and
that of GSH reductase, which catalyzes the reconversion of GSSG to GSH
are elevated in hippocampus and inferior parietal lobule of AD patients
(Aksenov et al., 1998; Lovell et al., 1998). This might reflect the protective
gene response to the increased peroxidation in the brain regions showing
severe AD pathology (Aksenov et al., 1998; Tchantchou et al., 2004a).
The levels of glutathione S-transferase, a protective enzyme against aldehydes
and especially 4-hydroxynonenal (HNE, a marker of lipid peroxidation) are
decreased in the brain and ventricular CSF of autopsied AD (Lovell et al.,
1998). APOE
/
mice maintained on folate-deficient diet demonstrate
similar increase in the activity of glutathione peroxidase and glutathione
reductase than that of APOE
/
mice on the complete diet. By contrast,
but consistent with observations made in AD patient brains, APOE
/
mice
display a significant decrease in glutathione S-transferase activity. The decrease
might be due to the methylation status of this enzyme. The similar increase in
GPx and GRactivity, which contributes in recycling the oxidized glutathione
(GSSG) back to the reduced form (GSH), combined with the significant
decrease in GST activity constitutes a justification for the increase GSH levels
in mice brain under oxidative damage (Shea and Rogers, 2002b; Tchantchou
et al., 2004a). The supplementation of APOE
/
mice with a potent methyl
donor, S-adenosylmethionine, when maintained on folate-deficient diet,
restores GST, GPx, and GR activity (Tchantchou et al., 2006b). This high-
lights the importance of potent methyl donors in the regulation of enzymes
that catalyze reactions involving the utilization of glutathione.
Considering the direct correlation between folate deprivation and
increase homocysteine levels, which exerts its neurotocixity via several
frameworks that include its ability to trigger increase b-amyloid deposition,
free radical formation, or its direct interaction with the plasma membrane,
combinatorial therapeutic approaches (McCaddon, 2006; Tchantchou et al.,
2004b) aiming at preventing homocysteine accumulation while maintaining
a normal methylation status provide a real hope to the management of
AD onset and at least part of the diseases symptoms.
Folate Deprivation, the Methionine Cycle, and Alzheimers Disease 93
REFERENCES
Aksenov, M. Y., Tucker, H. M., Nair, P., Aksenova, M. V., Butterfield, D. A., Estus, S., and
Markesbery, W. R. (1998). The expression of key oxidative stress-handling genes in
different brain regions in Alzheimers disease. J. Mol. Neurosci. 11, 151164.
Al-Gazali, L. I., et al. (2001). Abnormal folate metabolism and genetic polymorphism of the
folate pathway in a child with Down syndrome and neural tube defect. Am. J. Med.
Genet. 103, 28132.
Anonymous, L. I. (1998). Homocysteine lowering trialists collaboration. Lowering blood
homocysteine with folic acid based supplements: Meta-analysis of randomised trials. BJM
316, 894898.
Bottiglieri, T., Crellin, R., and Reynolds, E. H. (1995). Folate and neuropsychiatry.
In Folate in Health and Disease (L. B. Bailey, ed.), pp. 435462. Marcel Dekker,
New York.
Bronstrup, A., Hages, M., Prinz-Langenohl, R., and Pietrzik, K. (1998). Effects of folic acid
and combinations of folic acid and vitamin B12 on plasma homocysteine concentrations
in healthy young women. Am. J. Clin. Nutr. 68, 11041110.
Brotto, L. D., and Yang, Q. (2000). 5,10-Methylenetetrahydrofolate reductase gene variants
and congenital anomalies: A HuGE review. Am. J. Epidemiol. 151, 862867.
Cantoni, G. L. (1986). The centrality of S-adenosylhomocysteine in the regulation of the
biological utilization of S-adenosylmethionine. In Biological Methylation and Drug
Design, Experimental and Clinical Roles of SAM (C. Borchardt and P. M. Ueland,
eds.), pp. 227237. Humana Press.
Ceballos-Picot, I., Witko-Sarsat, V., Merad-Boudia, M., Nguyen, A. T., Thevenin, M.,
Jaudon, M. C., Zingraff, J., Verger, C., Jungers, P., and Descamps-Latscha, B. (1996).
Glutathione antioxidant system as a marker of oxidative stress in chronic renal failure.
Free Radic. Biol. Med. 21(6), 845853.
Chan, A., Tchantchou, F., Graves, V., Rozen, R., and Shea, T. B. (2006). Dietary
and genetic compromise in folate availability reduces acetylcholine and cognitive
performance: Critical role of S-adenosyl methionine. J. Health Nutr. Aging (in press).
Chen, J., Kyte, C., Valcin, M., Chan, W., Wetmur, J. G., Selhub, J., Hunter, D. J., and
Jing, M. A. (2004). Polymorphisms in the one-carbon metabolic pathway, plasma folate
levels and colorectal cancer in a prospective study. Int. J. Cancer 110(4), 617620.
Devlin, A. M., Arning, E., Bottiglieri, T., Faraci, F. M., Rozen, R., and Lentz, S. R. (2004).
Effect of mthfr genotype on diet-induced hyperhomocysteinemia and vascular function
in mice. Blood 103, 26242629.
Farkas, E., Jong, G. I., Apro, E., De Vos, R. A., Steur, E. N., and Luiten, P. G. (2000).
Similar ultrastructural breakdown of cerebrocortical capillaries in Alzheimers disease,
Parkinsons disease, and experimental hypertension. What is the functional link? Ann.
N. Y. Acad. Sci. 903, 7282.
Fisher, M. C., Zeisel, S. H., Mei-Heng, M., and Sadler, T. W. (2002). Perturbations in
choline metabolism cause neural tube defects in mouse embryos in vitro. FASEB J. 16,
619621.
Fuso, A., Seminara, L., Cavallaro, R. A., DAnselmi, F., and Scarpa, S. (2005). S-adeno-
sylmethionine/homocysteine cycle alterations modify DNA methylation status with
consequent deregulation of PS1 and BACE and beta-amyloid production. Mol. Cell.
Neurosci. 28, 195204.
Gilgun-Sherki, Y., Melamed, E., and Offen, D. (2001). Oxidative stress induced-
neurodegenerative diseases: The need for antioxidants that penetrate the blood brain
barrier. Neuropharmacology 40(8), 959975.
Holcomb, L., Gordon, M. N., McGowan, E., Yu, X., Benkovic, S., Jantzen, P., Wright, K.,
Saad, I., Mueller, R., Morgan, D., Sanders, S., Zehr, C., et al. (1998). Accelerated
94 Flaubert Tchantchou and Thomas B. Shea
Alzheimer-type phenotype in transgenic mice carrying both mutant amyloid precursor
protein and presenilin 1 transgenes. Nat. Med. 1, 97100.
Ho, P. I., et al. (2001). Homocysteine potentiates beta-amyloid neurotoxicity, role of
oxidative stress. J. Neurochem. 78, 249253.
Ho, P. I., Ashline, D., Dhitavat, S., Ortiz, D., Collins, S. C., Shea, T. B., and Rogers, E.
(2003). Folate deprivation induces neurodegeneration: Roles of oxidative stress and
increased homocysteine. Neurobiol. Dis. 14(10), 3242.
Huang, G. S., Yang, S. M., Hong, M. Y., Yang, P. C., and Liu, Y. C. (2000). Differential
gene expression of livers from ApoE-deficient mice. Life Sci. 68, 1928.
Kim, J.-M., Lee, H., and Chang, N. (2002). Hyperhomocysteinemia due to short-term folate
deprivationis relatedtoelectronmicroscopicchanges intherat brain. J. Nutr. 132, 34183421.
Kruman, I. I., Culmsee, C., Chan, S. L., Kruman, Y., Guo, Z., Penix, L., and
Mattson, M. P. (2000). Homocysteine elicits a DNA damage response in neurons that
promotes apoptosis and hypersensitivity to excitotoxicity. J. Neurosci. 20, 69206926.
Kruman, I. I., Kumaravel, T. S., Lohani, A., et al. (2002). Folic acid deficiency and
homocysteine impair DNA repair in hippocampal neurons and sensitize them to amyloid
toxicity in experimental models of Alzheimers disease. J. Neurosci. 22, 17521762.
Larsson, A., Orrenius, S., Holmgren, A., and Mannervik, B. (1983). Functions of
Glutathione. Biochemical Physiological Toxicological and Clinical Aspects,
pp. 14271434. Raven Press, New York.
Loehrer, F., Angst, C., Haefeli, W., et al. (1996). Low whole-blood S-adenosylmethionine
and correlation between 5-methyltetrahydrofolate and homocysteine in coronary artery
disease. Arterioscler. Thromb. Vasc. Biol. 16, 227233.
Loscalzo, J. (1996). The oxidant stress of hyperhomocysteinemia. J. Clin. Invest. 98, 57.
Lovell, M. A., Xie, C., and Markesbery, W. R. (1998). Decreased glutathione transferase
activity in brain and ventricular fluid in Alzheimers disease. Neurology 51, 15621566.
Malaguarnera, M., Ferri, R., Bella, R., Alagona, G., Carnemolla, A., and Pennisi, G. (2004).
Homocysteine, vitamin B12 and folate in vascular dementia and in Alzheimer disease.
Clin. Chem. Lab. Med. 42, 10321035.
Mattson, M. P., and Haberman, F. (2003). Folate and homocysteine metabolism therapeutic
targets in cardiovascular and neurodegenerative disorders. Curr. Med. Chem. 10(19),
19231929.
Mattson, M. P., and Shea, T. B. (2003). Folate and homocysteine metabolism in neural
plasticity and neurodegenerative disorders. Trends Neurosci. 26(3), 137146.
Mattson, M. P., Chan, S. L., and Duan, W. (2002). Modification of brain aging and
neurodegenerative disorders by genes, diet, and behavior. Physiol. Rev. 82, 637672.
McCaddon, A. (2006). Homocysteine and cognitive impairment; a case series in a General
Practice setting. Nutr. J. 5, 6.
McMahon, A. J., Green, J. T., Skeaff, M. C., Knight, G. R., Mann, I. J., and Williams, M. S.
(2006). A controlled trial of homocysteine lowering and cognitive performance.. N.
Engl. J. Med. 354, 27642772.
Miller, J. W. (1999). Homocysteine and Alzheimers disease. Nutr. Rev. 57, 126129.
Miller, A. L., and Kelly, G. S. (1997). Homocysteine metabolism: Nutritional modulation
and impact on health and disease. Altern. Med. Rev. 2, 234254.
Morrison, L. D. (1996). Brain S-adenosylmethionine levels are severely decreased in
Alzheimers disease. J. Neurochem. 67, 13281331.
Mosharov, E., Cranford, M. R., and Banerjee, R. (2000). The quantitatively important
relationship between homocysteine metabolism and glutathione synthesis by the trans-
sulfuration pathway and its regulation by redox changes. Biochemistry 39(42),
1300513011.
Mudd, S. H., Levy, H. L., and Skovby, F. (1995). Disorders of transsulfuration. In The
Metabolic and Molecular Bases of Inherited Disease (C. R. Scriver, A. L. W. S. Beaudet
Sly, and D. Valle, eds.), 7th Ed. pp. 12791327. McGraw-Hill, Inc.
Folate Deprivation, the Methionine Cycle, and Alzheimers Disease 95
OSuilleabhain, P. E., Sung, V., Hernandez, C., Lacritz, L., Dewey, R. B., Jr.,
Bottiglieri, T., and Diaz-Arrastia, R. (2004). Elevated plasma homocysteine level in
patients with Parkinson disease: Motor, affective, and cognitive associations. Arch. Neurol.
61, 865868.
Pacheco-Quinto, J., Rodriguez de Turco, E. B., DeRosa, S., Howard, A., Cruz Sanchez, F.,
Sambamurti, K., Refolo, L., Petanceska, S., and Pappolla, M. A. (2006). Hyperhomo-
cysteinemic Alzheimers mouse model of amyloidosis shows increased brain amyloid b
peptide levels. Neurobiol. Dis. 22(3), 651656.
Parihar, M. S., and Hemnani, T. (2004). Alzheimers disease pathogenesis and therapeutic
interventions. J. Clin. Neurosci. 11, 456467.
Raber, J., Wong, D., Buttini, M., Orth, M., Bellosta, S., Pitas, R. E., Mahley, R. W., and
Mucke, L. (1998). Isoform-specific effects of human apolipoprotein E on brain function
revealed in ApoE knockout mice: Increased susceptibility of females. Neurobiol. Proc.
Natl. Acad. Sci. USA 95(18), 1091410919.
Ramassamy, C., Averill, D., Beffert, U., Bastianetto, S., Theroux, L., Lussier-Cacan, S.,
Cohn, J. S., Christen, Y., Davignon, J., Quirion, R., and Poirier, J. (1999). Oxidative
damage and protection by antioxidants in the frontal cortex of Alzheimers disease is
related to the apolipoprotein E genotype. Free Radic. Biol. Med. 27(56), 544553.
Reynolds, E. H. (2002). Benefits and risks of folic acid to the nervous system. J. Neurol.
Neurosurg. Psychiatr. 72, 567571.
Sachdev, P. S., Valenzuela, M. J., Brodaty, H., Wang, X. L., Looi, J., Lorentz, L.,
Howard, L., Jones, M., Zagami, A. S., Gillies, D., et al. (2003). Homocysteine as a risk
factor for cognitive impairment in stroke patients. Dement. Geriatr. Cogn. Disord. 15,
155162.
Scarpa, S., Fuso, A., DAnselmi, F., and Cavallaro, R. A. (2003). Presenilin 1 gene silencing
by S-adenosylmethionine. FEBS Lett. 541, 145148.
Schulz, J. B., Lindenau, J., Seyfried, J., and Dichgans, J. (2000). Glutathione, oxidative stress
and neurodegeneration. Eur. J. Biochem. 267, 49044911.
Selhub, J. (1999). Homocysteine metabolism. Annu. Rev. Nutr. 19, 217246.
Selhub, J., and Miller, J. W. (1992). The pathogenesis of homocysteinemia, interruption of
the coordinate regulation by S-adenosylmethionine of the remethylation and transsul-
furation of homocysteine. Am. J. Clin. Nutr. 55, 131138.
Seshadri, S., Beiser, A., Selhub, J., Jacques, P. F., Rosenberg, H., DAgostino, R. B.,
Wilson, P. W. F., and Wolf, P. A. (2002). Plasma homocysteine as a risk factor for
dementia and Alzheimers disease. N. Engl. J. Med. 346, 476483.
Shea, T. B., and Rogers, E. (2002a). Homocysteine as a risk factor for Alzheimers disease.
N. Engl. J. Med. 25, 2007.
Shea, T. B., and Rogers, E. (2002b). Folate quenches oxidative damage in brains of
apolipoprotein E-deficient mice: Augmentation by vitamin E. Mol. Brain Res. 108, 16.
Shea, T. B., Lyons-Wieler, J., and Rogers, E. (2001). Homocysteine, folate deprivation and
Alzheimer neuropathology. J. Alzheimers Dis. 3, 17.
Shea, T. B., Rogers, E., Ortiz, D., and Sheu, M.-S. (2002). Apolipoprotein E deficiency
promotes increased oxidative stress and compensatory increases in antioxidants in brain
tissue. Free Radic. Biol. Med. 33, 11151120.
Shields, D. C., et al. (1999). The thermolabile variant of methylenetetrahydrofolate
reductase and neural tube defects: An evaluation of genetic risk and the relative impor-
tance of the genotypes of the embryo and the mother. Am. J. Hum. Genet. 64,
10451055.
Storch, K. J., Wagner, D. A., Burke, J. F., and Young, V. R. (1988). Quantitative study
in vivo of methionine cycle in humans using [methyl-2h3]- and [1-13c]methionine.
Am. J. Physiol. 255, E322E331.
96 Flaubert Tchantchou and Thomas B. Shea
Tchantchou, F., Graves, M., Ashline, D., Moran, A., Pimenta, A., Ortiz, D., Rogers, E., and
Shea, T. B. (2004a). Increased transcription and activity of glutathione synthase in
response to deficiencies in folate, vitamin E and apolipoprotein E. J. Neurosci. Res. 75,
508515.
Tchantchou, F., Graves, M., Ortiz, D., Rogers, E., and Shea, T. B. (2004b). Dietary
supplementation with 3-deaza adenosine, N-acetyl cysteine, and S-adenosyl methionine
provide neuroprotection against multiple consequences of vitamin deficiency and oxida-
tive challenge: Relevance to age-related neurodegeneration. Neuromol. Med. 6(23),
93104.
Tchantchou, F., Graves, M., Rogers, E., Ortiz, D., and Shea, T. B. (2005). N-acetyl
cysteine alleviates oxidative damage to central nervous system of ApoE-deficient mice
following folate and vitamin E-deficiency. J. Alzheimers Dis. 7(2), 135138.
Tchantchou, F., Graves, M., Ortiz, D., Rogers, E., and Shea, T. B. (2006a). Expression and
activity of methionine cycle genes are altered following folate and vitamin E deficiency:
Differential compensatory responses of normal and apolipoprotein E-deficient mice.
Nutr. Neurosci. 9(12), 1724.
Tchantchou, F., Graves, M., Ortiz, D., Chan, D., Rogers, E., and Shea, T. B. (2006b).
S-adenosyl methionine: A connection between nutritional and genetic risk factors for
neurodegeneration in Alzheimers disease. J. Health Nutr. Aging (in press).
Ulrich, C. M., Robien, K., and Sparks, R. (2002). Pharmacogenetics and folate metabolism:
A promising direction. Pharmacogenomics 3(3), 299313.
Ureno, M., Haruhiko, S., Kenji, K., Masayuki, O., Ichiro, A., and Masanori, H. (2001).
Ultrastructural and permeability features of microvessels in the hippocampus, cerebellum
and pons of senescence-accelerated mice. Neurobiol. Aging 22, 469478.
Watanabe, M., Osada, J., Aratani, Y., Kluckman, K., Reddick, R., Malinow, M. R., and
Maeda, N. (1995). Mice deficient in cystathionine beta-synthase: Animal models for mild
and severe homocyst(e)inemia. Proc. Natl. Acad. Sci. USA 92(5), 15851589.
Wolters, M., Strohle, A., and Hahn, A. (2004). Cobalamin: A critical vitamin in the elderly.
Prev. Med. 39, 12561266.
Folate Deprivation, the Methionine Cycle, and Alzheimers Disease 97
C H A P T E R F O U R
Molecular Mechanisms of Adaptation
to Folate Deficiency
Ilan Ifergan* and Yehuda G. Assaraf
Contents
I. Folate Metabolism 101
II. Pathological States Associated with Folate Deficiency or
Nutritional Folate Insufficiency 105
A. Folate deficiency, neural tube defects and congenital
heart defects 105
B. Folate deficiency, homocysteinemia, and atherosclerotic
cardiovascular disease 106
III. Molecular Mechanisms of Adaptation to Folate Deprivation 108
A. The role of folate-dependent enzymes in adaptation
to folate deficiency 108
B. Cellular retention of folates: The key role of polyglutamylation 110
C. Overexpression of folate influx systems 112
D. Downregulation of folate efflux systems 123
References 131
Abstract
Folic acid is an essential vitamin for a wide spectrum of biochemical reactions;
however, unlike bacteria and plants, mammals are devoid of folate biosynthesis
and thus must obtain this cofactor from exogenous sources. Therefore, folate
deficiency may impair the de novo biosynthesis of purines and thymidylate and
thereby disrupt DNA and RNA metabolism, homocysteine remethylation, methi-
onine biosynthesis, and subsequent formation of S-adenosylmethionine (the
universal methyl donor) which in turn may lead to altered methylation reac-
tions. This impaired folate-dependent intracellular metabolism can lead to
several key pathologies including, for example, megaloblastic anemia, homo-
cysteinemia, cardiovascular disease, embryonic defects, in particular neural
tube defects (NTDs), congenital heart defects, and possibly cancer. The current
review presents and evaluates the up-to-date knowledge regarding the
Vitamins and Hormones, Volume 79
#
2008 Elsevier Inc.
ISSN 0083-6729, DOI: 10.1016/S0083-6729(08)00404-4 All rights reserved.
* The Fred Wyszkowski Cancer Research Laboratory, Department of Biology, Technion-Israel Institute of
Technology, Haifa 32000, Israel
{
To whom correspondence should be addressed; Email: assaraf@tx.technion.ac.il
99
molecular mechanisms underlying cellular survival and/or adaptation to folate
deficiency or insufficiency. These mechanisms of adaptation to folate deficiency
generally associated with folate uptake, intracellular folate retention, folate-
dependent metabolism, and active folate efflux specifically include: (a) Up- or
downregulation of various folate-dependent enzymes like dihydrofolate reduc-
tase (DHFR) and thymidylate synthase (TS), (b) Cellular retention of folates via
polyglutamylation by the enzyme folylpoly-g-glutamate synthetase (FPGS),
(c) Overexpression of folate influx systems including the reduced folate carrier
(RFC), folate receptor (FR) as well as the proton-coupled folate transporter
(PCFT), a recently identified intestinal folate influx transporter optimally func-
tioning at the acidic microclimate of the upper intestinal epithelium, (d) Down-
regulation of ATP-driven folate efflux transporters of the multidrug resistance
protein (MRP; ABCC) family and breast cancer resistance protein (BCRP; ABCG2)
that belong to the multidrug resistance (MDR) efflux transporters of the ATP-
binding cassette (ABC) superfamily. Moreover, the intricate interplay between
various components of the adaptive response to folate deprivation is also
discussed. 2008 Elsevier Inc.
Abbreviations
ABC ATP-binding cassette
AICARTF aminoimidazole carboxamide ribonucleotide
transformylase
ALL acute lymphoblastic leukemia
BCRP breast cancer resistance protein
cDNA complementary DNA
CFD cerebral folate deficiency
CHO Chinese hamster ovary
CNS central nervous system
CSF cerebrospinal fluid
DHF dihydrofolate
DHFR dihydrofolate reductase
FPGS folylpoly-g-glutamate synthetase
FR folate receptor
GARTF glycinamide ribonucleotide transformylase
GGH g-glutamyl hydrolase
GSH glutathione
HFM hereditary folate malabsorption
RFC reduced folate carrier
IMP inosine monophosphate
LF low folate
MDR multidrug resistance
100 Ilan Ifergan and Yehuda G. Assaraf
MFS major facilitator superfamily
MRPs multidrug resistance proteins
MSDs membrane-spanning domains
MTHFR methylene-tetrahydrofolate reductase
MTX methotrexate
NBF nucleotide-binding fold
NTD neural tube defect
PABA p-amino benzoic acid
pB promoter B
PCFT proton-coupled folate transporter
Pgp P-glycoprotein
PI3K phosphoinositol-3-kinase
ROS reactive oxygen species
SAM S-adenosylmethionine
SLC solute carrier family
SHMT serine hydroxymethyltransferase
THF tetrahydrofolate
TMD transmembrane domain
TS thymidylate synthase
I. Folate Metabolism
Folic acid, the oxidized form of vitamin B
9
, and its reduced derivative,
tetrahydrofolate (THF; Fig. 4.1), are water-soluble vitamins commonly
termed folates. The term folate stems from the Latin word folium which
means leaf; indeed, folates are present in substantial amounts in green leafy
vegetables. THF cofactors play an essential role as one-carbon donors and
acceptors in several crucial intracellular metabolic reactions. Folic acid is
composed of three covalently linked components: a pteridine ring, p-amino
benzoic acid (PABA) and a glutamate residue (Fig. 4.1). Mammals are able
to synthesize the pteridine ring. However, unlike bacteria and plants,
mammals are devoid of the enzymatic capacity of coupling pteridine to
PABA (Birn, 2006) and thus are absolutely dependent on folate uptake from
exogenous sources. These sources include intestinal uptake of folates bio-
synthesized by intestinal bacterial flora as well as dietary folate intake
particularly from green leafy vegetables, fruits, grain, yeast, liver, and dairy
products (Birn, 2006).
Biologically active folates exist predominantly in the reduced form (i.e.,
the two double bonds on the pteridine ring are enzymatically reduced),
thereby resulting in THF (Adams et al., 1970); Figs. 4.1 and 4.2). Inside the
cell, THF cofactors function as both acceptors and donors of one-carbon
Adaptation to Folate Deprivation 101
5-CH
3
-THF
N
N
O
NH CH C
COOH OH
NH
C
O
OH
CH
3
H
N
N N
N
H
N
N N
N
N
N
Folic acid
O
NH CH
2
CH
2
CH
2
CH
2
CH
2
CH
2
NH
2
CH
3
CH
3
CH
2
CH
2
CH
2
CH
2
CH
2
CH
2
CH
2
CH
2
CH C
COOH OH
NH
H
2
N
H
2
N
H
2
N
H
2
N
H
2
N
C
O
OH
N
N
N
N
N N
O
NH CH C
COOH
N
C
O
OH
N
N N
N
1
2
3
4
5
6
7
8
5,10-CH
2
-THF
O
NH CH C
COOH
OH N
C
O
OH
N
N
H
CH
2
C
N
N
N N
N
N
O
NH CH C
COOH OH
NH
C
O
OH
5-CHO-THF
(leucovorin)
CHO
H
N
N
N
N
CHO
CH
2
N
N
Methotrexate
Chemical structures Chemical category
Oxidized folate
Reduced folate
Reduced folate
Reduced folate
Antifolate
(DHFR inhibitor)
Pteridine ring
p-amino benzoic acid Glutamyl residue
CH
2
Figure 4.1 Chemical structures of oxidized and reduced folates as well as antifolates.
The one-carbon units donated during purine and thymidylate biosynthesis are
highlighted in gray.
102 Ilan Ifergan and Yehuda G. Assaraf
units. These one-carbon units are covalently linked to the THF cofactor
and exist as a wide range of oxidized and reduced forms of the one-carbon
moiety, including methyl, formyl and methylene derivatives (Figs. 4.1 and
4.2). The primary source of one-carbon units for mammalian folate metab-
olism is first donated from serine to the pteridine ring of THF thereby
resulting in the formation of glycine and 5,10-methylene-THF, respectively
(Fig. 4.2). The various THF cofactors include 5,10-methylene-THF,
10-formyl-THF, and 5-methyl-THF (the major folate form in the plasma)
serve as one-carbon donors in various biochemical reactions (Carmel et al.,
2003; Qureshi et al., 1994; Scott, 1999) as depicted in Fig. 4.2; specifically,
cytosolic methionine synthase transfers a methyl group from 5-methyl-THF
to homocysteine in a methyl-cobalamin-dependent reaction, thereby result-
ing in the formation of methionine (Fig. 4.2). The latter is further converted
to S-adenosylmethionine (SAM). In the methylation cycle which occurs in
all nucleated cells, SAM serves as the key methyl group donor of most
biological methylation reactions including that of CpG island DNA methyl-
ation that regulates gene expression (Kim, 2004, 2005), protein methylation
AICARTF
PRPP
GMP AMP
IMP
DHF
TS
THF
THF THF
dTMP
dUMP
Hcy
Met
Glycine
Serine
Folic
acid
Cytoplasm
GARTF
10 -CHO-THF
5,10-CH
2
-THF 5-CH
3
-THF
DNA
5 -CHO-THF
SAM
X
CH
3
X
DHFR
SHMT
MTHFR
DHFR
MS
Figure 4.2 Simplified scheme of intracellular folate metabolism in mammalian cells
with emphasis on the de novo biosynthesis of purines and thymidylate. Abbreviations
of the enzymes and intermediate compounds that are involved in these pathways:
AICARTF, aminoimidazole carboxamide ribonucleotide transformylase; DHF, dihy-
drofolate; DHFR, dihydrofolate reductase; FPGS, folylpoly-g-glutamate synthetase;
GARTF, glycinamide ribonucleotide transformylase; GGH, g-glutamyl hydrolase;
Hcy, homocysteine; IMP, inosine monophosphate; Met, methionine; MS, methionine
synthase; MTHFR, methylene-tetrahydrofolate reductase; PRPP, Phosphoribosyl
pyrophosphate; SAM, S-adenosylmethionine; SHMT, serine hydroxymethyltransfer-
ase; THF, tetrahydrofolate; TS, thymidylate synthase.
Adaptation to Folate Deprivation 103
which serves as an important posttranslational modification, and lipid meth-
ylation important for their biosynthesis including that of phosphatidylcholine
(Stokstad) (Fig. 4.2). Moreover, methylation is a crucial step in the metabo-
lism of neurotransmitters and detoxification of xenobiotics. Hence, folate
deficiency may lead to elevated homocysteine levels due to decreased utili-
zation of this amino acid as well as to altered intracellular methylation
processes. In addition to the above-mentioned role of THF cofactors in
the biosynthesis of glycine and methionine along with the metabolism of
serine and homocysteine, respectively, THF coenzymes play a key role in the
catabolism of histidine and formic acid. Importantly, THF coenzymes are
essential for cellular proliferation due to their primary role in the de novo
biosynthesis of purines and thymidylate that are essential for DNAreplication
(Fig. 4.2). Several key enzymes use reduced folates as cofactors in these
biosynthetic pathways (Fig. 4.2). Specifically, thymidylate synthase (TS)
catalyzes the conversion of dUMP to dTMP through the transfer of a methyl
group from 5,10-methylene-THF to dUMP. Similarly, the cofactor
10-formyl-THF contributes one-carbon units in each of two key de novo
biosynthetic transformylase reactions of purine metabolism (Fig. 4.2). The
first reaction is catalyzed by the enzyme glycinamide ribonucleotide trans-
formylase (GARTF) which transfers a formyl group from 10-formyl-THF
resulting in the formation of the imidazole ring of purines. Whereas, a more
downstreamreaction involves an additional formyl transferase reaction that is
carried out by aminoimidazole carboxamide ribonucleotide transformylase
(AICARTF), resulting in the formation of the purine intermediate inosinic
acid which is commonly termed inosine monophosphate (IMP) (Fig. 4.2).
The latter serves as a crucial intermediate in a biochemical Y-junction which
leads to the regulated and balanced formation of the purine nucleotides AMP
and GMP. Importantly, the folate cofactor is utilized in a cyclic manner
as follows: dihydrofolate (DHF), the by-product of the TS-dependent
reaction, is recycled to the biologically active THF by the enzyme dihydro-
folate reductase (DHFR) in an nicotinamide adenine dinucleotide
phosphate (NADPH)-dependent reduction. The latter THF as well as the
THF by-product of the purine metabolismcan be further converted to 5,10-
methylene-THF and 10-formyl-THF. The latter two folate cofactors are
reused in the de novo biosynthesis of purines and thymidylate precursors of
nucleic acids as has been previously described (Fig. 4.2).
On the basis of plethora of folate-dependent metabolic processes, a
deficiency in intracellular THF pools may therefore impair nucleotide
biosynthesis, homocysteine remethylation, methionine biosynthesis, as
well as cellular methylation processes. This altered intracellular biochemical
status can lead to several disorders including megaloblastic anemia, leuco-
penia and thrombocytopenia, homocysteinemia, cardiovascular disease,
embryonic defects, in particular neural tube defects (NTDs) and congenital
heart defects and possibly cancer (Stanger, 2002). These pathological states
104 Ilan Ifergan and Yehuda G. Assaraf
associated with folate deficiency are discussed in greater detail in the
following chapters. Moreover, a similar effect of folate deprivation can be
achieved by using folic acid antagonists known as antifolates that inhibit the
biosynthesis of purines and/or thymidylate. This rationale has been
exploited for the introduction of a wide spectrum of antifolate-based
therapies of various diseases that are characterized by an abnormal cellular
proliferation including cancer, rheumatoid arthritis, and psoriasis ( Jukes,
1987). Methotrexate (MTX) (Fig. 4.1), the first and most studied antifolate,
binds stoichiometrically and tightly (K
D
1 pM) to the target enzyme
DHFR with an affinity that is a million fold higher than the natural
substrate, DHF (K
m
1 mM) (Bertino, 1993; Ozaki et al., 1981). However,
at least 95% of intracellular DHFR must be inhibited by MTX in order to
block DNA replication (Jackson and Harrap, 1973). In addition to inhibi-
tion of DNA synthesis, antifolate treatment may result in misincorporation
of 2
0
-deoxyuridine 5
0
-triphosphate (dUTP) into DNA, thereby leading to
DNA strand breaks and cell death (Richards et al., 1986). MTX is currently
used for the combination chemotherapeutic treatment of various human
malignancies including childhood acute lymphoblastic leukemia (ALL),
non-Hodgkins lymphoma, osteosarcoma, head and neck cancer, choriocar-
cinoma, small cell lung cancer, and breast cancer (Bertino, 1993; Schornagel
and McVie, 1983). A detailed discussion of the currently available antifolates
is beyond the scope of this review; however, this has been the subject of
a recent review from this laboratory (Assaraf, 2007). Indeed, several key
folate-dependent enzymes have been successfully targeted by one or various
antifolates that are widely used in the combination chemotherapeutic treat-
ment regimen of various human cancers of hematolymphoid lineage or
epithelial origin (Walling, 2006).
II. Pathological States Associated with
Folate Deficiency or Nutritional
Folate Insufficiency
A. Folate deficiency, neural tube defects and congenital
heart defects
The closure of the neural tube occurs during early stages of embryogenesis.
This developmental process heavily relies on the interactions between
genetic, environmental, and nutritional factors. Failure of neural tube
closure is a common congenital malformation resulting in significant mor-
bidity and mortality. NTDs are complex congenital malformations of the
central nervous system (CNS) and the most common (with a prevalence of
1:1000 births) and severe forms of NTDs include anencephaly and spina
bifida (Blom et al., 2006). Anencephaly is an NTD that occurs when there is
Adaptation to Folate Deprivation 105
failure of neurulation, in which the neural folds do not fuse at the cranial
end of the developing embryo, and hence, these infants lack most or all of
the brain tissue. Infants with anencephaly are stillborn or die shortly after
birth. Spina bifida is a posterior NTD that is caused by the failure of the
neural tube to fuse at the caudal end. Consequently, infants with Spina
bifida can have an open lesion on their spine, where significant damage to
the nerves and spinal cord has occurred or the lesion can be closed. Infants
with spina bifida can survive but are at high risk of developing psychosocial
maladjustments. Maternal nutrition factors contribute significantly to the
various etiologies of NTDs. In a seminal study reported on 1976, Smithells
and colleagues (Smithells et al., 1976) discovered that the diets and postpar-
tum blood of women who had given birth to a fetus with an NTD were
deficient in several micronutrients, particularly folic acid. Over the past two
decades, it has been well established that women can reduce their risk of
having an NTD-affected pregnancy by as much as 5075% by taking folic
acid supplementation. Small nonrandomized studies in women who previ-
ously had NTD-affected pregnancies revealed that intake of folic acid supple-
ments during the periconceptional period resulted in a fourfold reduction in
the NTD recurrence risk (Laurence et al., 1981; Seller and Nevin, 1984;
Smithells et al., 1976, 1981, 1983; Vergel et al., 1990). Moreover, in a
thorough, double-blind, placebo-controlled, randomized trial performed by
the Medical Research Council in the UK it was confirmed that supplemen-
tation with a daily dose of 4 mg folic acid resulted in a beneficial effect of a
threefold reduction in the recurrence risk of NTDs (Group, 1991). However,
it is still unclear as to what are the molecular mechanisms by which folic acid
supplementation markedly decreases the risk of an infant being born with an
NTD or other congenital malformations. Moreover, it is completely unclear
why a significant fraction of women who do take folic acid supplementation
during the periconceptional period give birth to NTD-harboring infants.
B. Folate deficiency, homocysteinemia, and atherosclerotic
cardiovascular disease
As detailed above, reduced folates are cofactors that are essential for the
biosynthesis of purine and thymidine nucleotides. Furthermore, folates are
also necessary for the biosynthesis of methionine from homocysteine and
5-CH
3
-THF in a methyl-cobalamin (i.e., methyl-B
12
)-dependent reaction
catalyzed by methionine synthase (Fig. 4.2). Impairment of folate-mediated,
one-carbonmetabolic pathways canresult fromfolate deficiency and/or nutri-
tional folate insufficiency (i.e., a diet that is relatively poor in folates), single
nucleotide polymorphisms and inactivating mutations in folate-dependent
enzymes and folate transporters. These result in homocysteinemia and
increased risk of various pathologies including atherosclerotic cardiovascular
106 Ilan Ifergan and Yehuda G. Assaraf
disease and vascular thrombosis (Arnesen et al., 1995; Durand et al., 2001;
Graham et al., 1997; Rasouli et al., 2005; Ubbink et al., 1993). Data suggest
that folate deficiency was the most common vitamin deficiency in the United
States, thereby affecting 10%of the population; furthermore, more than50%
of the childrenandelderly that live inpoverty were folate deficient (Baileyet al.,
1982). These statistics of folate deficiency were found to be true also for other
non-Western countries including, for example, Thailand (Assantachai and
Lekhakula, 2007) and rural areas of India (Pathak et al., 2004). However,
sincethemandatedfortificationof processedgrains withfolic acidintheUnited
States and Canada in 1998, the incidence of folate deficiency in most popula-
tions in these countries has dramatically declined ( Joelson et al., 2007).
Epidemiological evidence suggests that hyperhomocysteinemia is an
independent single risk factor for arterial thrombotic diseases such as acute
myocardial infarction, stroke, peripheral ischemic occlusive disorders as
well as venous thromboembolism (Arnesen et al., 1995; Durand et al.,
2001; Graham et al., 1997; Rasouli et al., 2005). In this respect, panoply
of epidemiological studies has accumulated over the past decade; these
studies establish a tight association between homocysteinemia and the
significantly increased risk of atherosclerotic cardiovascular disease. It has
been well established that folate deficiency results in elevated levels of
serum homocysteine (i.e., homocysteinemia). Hence, in folate-deficient
individuals, decreased levels of THF cofactors (e.g., 5-CH
3
-THF) limit
metabolic flux through the methionine synthase reaction, with conse-
quent accumulation of homocysteine, the substrate of this enzyme.
Importantly however, dietary folate supplementation can readily normalize
plasma homocysteine levels and may thereby reduce the risk of coronary
artery disease (Graham and OCallaghan, 2000). Studies with animal models
of atherosclerosis and thrombosis have provided evidence that even mod-
erate elevation of homocysteine levels can produce damage to vascular
endothelium and can enhance platelet aggregation (Sauls et al., 2004,
2007). These vasotoxic and other deleterious effects of homocysteine
are believed to result from the following characteristics (Ramakrishnan
et al., 2006):
(a) During the oxidation of excess homocysteine to homocystine and
disulphides, reactive oxygen species (ROS) which are potent oxidants
are generated in the blood which can cause severe endothelial injury.
For example, endothelial cell injury due to copper-catalyzed hydrogen
peroxide formation from cytseine has been demonstrated in vitro
(Starkebaum and Harlan, 1986).
(b) Homocysteine can directly interact in a thiolation reaction by binding
to thiol groups of proteins and thereby form disulphides which can
markedly alter the function of various proteins.
Adaptation to Folate Deprivation 107
(c) Homocysteine can be converted to a highly reactive thiolactone which
could react with proteins resulting in the formation of NH-CO-
adducts, thereby altering the function of various proteins and enzymes.
For instance, a potential mechanism for the thrombotic tendency in
hyperhomocysteinemic patients has been recently identified; thus,
modification of fibrinogen by homocysteine thiolactone increased its
resistance to fibrinolysis (Sauls et al., 2006).
III. Molecular Mechanisms of Adaptation
to Folate Deprivation
This chapter reviews and discusses the up-to-date knowledge regard-
ing the molecular mechanisms that regulate folate homeostasis under folate
replete and deplete conditions. Our specific aim is to describe the role of
folate-dependent enzymes, folate polyglutamylation as well as folate influx
and efflux transport systems in cellular survival and/or adaptation to folate
deficiency and/or insufficiency.
A. The role of folate-dependent enzymes in adaptation
to folate deficiency
De novo biosynthesis of purines and thymidylate precursors of nucleic acids is
absolutely dependent on THF cofactors. Hence, cellular proliferation will be
eventually blocked under conditions of severe folate deficiency. One possi-
ble mechanism of adaptation to folate deprivation lies within the folate
metabolic pathway. Specifically, the DHF by-product is recycled to the
biologically active THF cofactor by the enzyme DHFR; the resultant
THF cofactor can be further converted to 5,10-methylene-THF and
10-formyl-THF for utilization in the de novo biosynthesis of purines and
thymidylate as has been previously described (Fig. 4.2). Hence, augmented
activity of DHFR should theoretically increase the recycling rate of DHF,
at least in some cells, and thus compensate, at least partially, for conditions in
which folate is deprived from the growth medium. Indeed, analysis of three
independent clonal variants of Chinese hamster lung fibroblasts including
FA3, FA7, and FA14 selected under low folate (LF) conditions [15 pM
5-formyl-THF; (Lamers et al., 2006)], revealed that each clone had a fivefold
to sixfold increase in both DHFR activity and protein levels (Zhu et al.,
2002). Similarly, DHFRgene amplification was documented after treatment
with MTX both in cultured cells (Schimke, 1984) and in clinical samples
from patients that were treated with this antifolate (Horns et al., 1984; Trent
et al., 1984). For certain concentrations of MTX, the significant increase in
the intracellular levels of DHFR produces sufficient free enzyme molecules
108 Ilan Ifergan and Yehuda G. Assaraf
to catalyze its biosynthetic reaction resulting in the production of THF.
Therefore, increased activity of DHFR plays a key role under conditions of
folate deficiency or DHFR inhibition by antifolates, thereby producing
more enzyme molecules to minimize the disruption of folate metabolism.
TS activity was also examined under conditions of folate deficiency; as
has been previously mentioned, TS catalyzes the conversion of dUMP
to dTMP by transferring a single methyl group from the cofactor 5,10-
methylene-THF to dUMP (Fig. 4.2). In this study, four colon cancer cell
lines (C26-A, C2610, C26-G, and WiDr) and three squamous cell carci-
noma lines of the head and neck (HNSCC) (11B, 14C, and 22B) were
adapted to grow in a culture medium containing LF levels (i.e., 0.52.5 nM
leucovorin) (Backus et al., 2000). This study revealed that whereas folate
deprivation was associated with up to threefold increased catalytic activity of
TS in squamous cell carcinoma of the head and neck, the folate-depleted
colon cancer cells showed up to sevenfold decreased catalytic activity of
TS when compared to their folate-replete parental controls. The increased
activity of TS in the folate-depleted squamous carcinoma cells of the head
and neck may compensate for the folate-deprived condition, thus facilitat-
ing an enhanced utilization of folates for the biosynthesis of dTMP. How-
ever, the decreased activity of TS in the folate-deprived colon cancer cells
may serve as an important mechanism for rebalancing the intracellular
concentration of nucleotides for cellular utilization. Specifically, the
decreased activity of TS in these colon cancer cells may release more THF
cofactors for de novo purine biosynthesis; moreover, decreased TS activity
may result in diminished levels of dTMP, the latter of which is presumably
less required as the proliferation rate is decreased due to the severe folate
deficiency. Indeed, folate deprivation of cultured cells has been reported to
increase the intracellular ratio of dUMP to dTMP by as high as tenfold
( James et al., 1994; Melnyk et al., 1999). Consistent with the hypothesis that
certain cells downregulate DNA synthesis under folate-deprived conditions,
folate-deficient human colon adenocarcinoma HCT116 cells appeared
to favor a preferential shuttling of the flux of one-carbon units to the
methionine cycle (upregulation of methylene-tetrahydrofolate reductase,
MTHFR) and thereby suppress the nucleotide biosynthetic pathway
(including downregulation of the enzymes SHMT, TS, and DHFR;
Fig. 4.2) (Hayashi et al., 2007). Therefore, suppression of the pathway of
nucleotide biosynthesis, at least partially, may be strategically exercised by
cells as an adaptation and/or survival mechanism under folate deprivation.
Additional studies are necessary to reveal the role of folate-dependent
enzymes in the complex intracellular biochemical network of folate-
dependent pathways under folate-deprived conditions. Moreover, compu-
terized simulations of de novo biosynthesis of purines and thymidylate
precursors of nucleic acids currently became available (Assaraf et al., 2006;
Seither et al., 1989); these computer-aided models may be used to analyze
Adaptation to Folate Deprivation 109
the nature of this important pathway under conditions of folate deficiency
and thereby predict the possible role of specific folate-dependent enzymes
in adaptation to conditions of folate deprivation.
B. Cellular retention of folates: The key role
of polyglutamylation
Whereas the monoglutamate formof 5-CH
3
-THF is the primary circulatory
folate, the polyglutamylated forms of 5-CH
3
-THF as well as other reduced
folates serve as the abundant intracellular THF cofactors. The enzyme folyl-
poly-g-glutamate synthetase (FPGS) which exists in both the cytosol and
mitochondria (McGuire et al., 2000), catalyzes the polyglutamylation of
folates and antifolates through the addition of multiple equivalents of gluta-
mate units tothe g-carboxyl residue of THFcofactors and certainhydrophilic
antifolates (Fig. 4.3). In contrast to FPGS, the enzyme g-glutamyl hydrolase
(GGH) catalyzes the hydrolysis of these terminal glutamate residues from
polyglutamylated folates and antifolates (Fig. 4.3) (Shane, 1995). Intracellular
folates exist mainly as poly-g-glutamate derivatives, with the parent THF
being elongated by 710 glutamyl residues (McGuire et al., 1980; Shane,
1989). Importantly, the long chain (n > 3) polyglutamylated (anti)folate
derivatives are no longer substrates of efflux systems such as multidrug
resistance proteins (MRPs) (Wielinga et al., 2005; Zeng et al., 2001), breast
cancer resistance protein (BCRP) (Volk and Schneider, 2003) as well as
bidirectional folate transporters including the reduced folate carrier (RFC)
(Matherly and Goldman, 2003). Hence, polyglutamylation is the primary
determinant of intracellular retention and sequestration of THFcofactors and
polyglutamatable antifolates. Moreover, folate polyglutamate congeners dis-
play higher affinities for various folate-dependent enzymes relative to their
non-polyglutamylated forms (Allegra et al., 1985, 1987). Therefore, folylpo-
lyglutamylation increases the efficiency of THF cofactor utilization. Addi-
tionally, a mitochondrial FPGS species exists which allows mitochondria to
accumulate folylpolyglutamates thereby resulting in the biosynthesis of gly-
cine (Lin et al., 1993). The FPGS-dependent entrapment of intracellular
THF cofactors suggests that increased cellular activity of this enzyme should
be highly beneficial to cells subjected to conditions of folate deprivation.
Several studies lend support to this hypothesis:
(a) An increased FPGS activity was documented in several tumors as well as
in the liver and kidney of mice that were subjected to a LF diet for two
weeks (Mendelsohn et al., 1996).
(b) Consistently, mice injected i.v. with a single dose of the antifolate
lometrexol (5,10-dideaza-tetrahydrofolate), an inhibitor of GARTF,
accumulated more drug in their livers when compared to untreated
controls. Moreover, analysis of antifolate polyglutamates revealed that
110 Ilan Ifergan and Yehuda G. Assaraf
longer polyglutamate forms of lometrexol (hepta- and octa- species of
lometrexol) appeared earlier and persisted longer in liver of the folate-
deprived mice when compared to the control group of mice
(Mendelsohn et al., 1996). The increased activity of FPGS may there-
fore enhance (anti)folate polyglutamylation and thereby improve intra-
cellular folate retention. Moreover, the markedly increased affinities of
various folate-dependent enzymes for polyglutamylated folate deriva-
tives relative to the non-glutamylated forms (Allegra et al., 1985, 1987)
should be particularly important under conditions of folate deprivation.
(c) An additional study aimed at understanding the molecular mechanisms
underlying adaptation of cultured breast cancer cells to states of
folate deficiency also supports the important role that increased polygluta-
mylationplays inadaptationtofolate deficiency(Iferganet al., 2004). Inthis
study, MCF-7breast carcinoma cells andtheir mitoxantrone(ananticancer
Folate(glu
1
)
NH
CH
2
CH
2
CH
COOH
C
O
OH
GGH
N
N
O
NH CH
2
CH
2
CH
2
CH C
COOH OH
NH
H
2
N
C
O
OH
H
CH
3
N
N
FPGS
NH CH
2
CH
2
CH
2
CH C
O COOH OH
NH
H
2
N
C
O
H
CH
3
N
N
N
N
Folate(glu
n
)
n
Glutamate
+
ATP
ADP+Pi
Figure 4.3 The reaction of folate polyglutamylation by FPGS and hydrolysis of folate
polyglutamates by GGH. Note that various reduced folate derivatives undergo these
reactions including 5-methyl-THF as depicted in this scheme.
Adaptation to Folate Deprivation 111
drug which is a specific BCRP substrate)-resistant MCF-7/MR subline,
with BCRP (an ATP-driven efflux transporter of mitoxantrone and
folates) overexpression were gradually deprived (3.5 months) of folic
acid from 2.3 mM to 3 nM resulting in the sublines MCF-7/LF and
MCF-7/MR-LF, respectively. Consistent with the above results, folate
deprivation resulted in statistically significant increases of 63% and 20% in
FPGS activity of MCF-7/MR-LF and MCF-7/LF cells, respectively,
relative totheir folate-replete counterparts (Iferganet al., 2004). Adecrease
in one or more folate efflux systems along with increased FPGS activity
were associated with an augmented ability of the LF-adapted sublines to
accumulate folates.
(d) Furthermore, a study with colonic epithelial cells aimed at determining
the alterations occurring at the transcript level of various genes involved
in intracellular folate metabolism and one-carbon transfer reactions
under folate-deprived conditions was recently published (Hayashi
et al., 2007). Consistent with previously mentioned studies, a real-
time quantitative RTPCR analysis revealed that the steady-state
mRNA levels of FPGS were 400% and 16% higher in folate-deficient
HCT116 and Caco2 human colon adenocarcinoma cells, respectively,
relative to their folate-replete counterparts (Hayashi et al., 2007). These
results indicate that transcriptional upregulation of FPGS and/or
decreased degradation of FPGS transcripts (i.e., increased mRNA sta-
bility) are plausible mechanisms of cellular adaptation to folate defi-
ciency. Taken in toto, these cumulative results strongly suggest that
increased FPGS activity may shift the intracellular equilibrium of folates
toward the long chain, hyperpolyglutamylated congeners, thereby
decreasing active folate export via various ATP-driven folate efflux
trasnporters; for a recent review see Assaraf (2006). Hence, increased
FPGS activity appears to play a major protective role and is apparently an
important survival factor in cellular adaptation to folate deficiency.
C. Overexpression of folate influx systems
1. Cellular uptake of folates and MTX
Whereas a balanced diet may contain both folate monoglutamates and
polyglutamates, the latter are enzymatically hydrolyzed to the monogluta-
mate form by glutamate carboxypeptidase II (GCPII), an exopeptidase
anchored to the intestinal apical brush border membrane (Chandler et al.,
1986). The monoglutamate form of 5-methyl-THF, additional reduced
folate derivatives as well as the oxidized folate form, folic acid, are divalent
anions that cannot traverse lipid bilayers by simple diffusion (Dembo and
Sirotnak, 1984; Goldman and Matherly, 1986; Sirotnak and Tolner, 1999).
Therefore, folate monoglutamate uptake into mammalian cells must be
facilitated by specialized carrier-mediated systems (Dembo and Sirotnak,
112 Ilan Ifergan and Yehuda G. Assaraf
1984; Goldman and Matherly, 1986; Ratnam and Freisheim, 1992)
(Figs. 4.4 and 4.5). Indeed, carrier-mediated systems for the uptake of folates
across biological membranes are necessary for intestinal folate absorption
and renal reabsorption as can be found in the upper small intestine and
proximal renal tubules, respectively, as well as for folate uptake into cells
located within the intact embryo and the various tissues of adult organisms
(Lee et al., 1992; Matherly and Goldman, 2003; Ratnam and Freisheim,
1992; Rubin et al., 1967; Selhub and Rosenberg, 1978, 1986; Sirotnak and
Tolner, 1999). Three specialized transport systems exist that can accommo-
date the transport of folates and antifolates across biological membranes
(Matherly and Goldman, 2003):
(i) The RFC is a major uptake route for the transport of folates into
mammalian cells (Matherly and Goldman, 2003) (Figs. 4.4 and 4.5).
RFC (SLC19A1) functions as a bidirectional anion exchanger
(Goldman, 1971; Henderson and Zevely, 1981) with a high affinity
(K
m
1 mM) for reduced folates and hydrophilic antifolates including
MTX (K
m
510 mM) but low affinity (K
m
200400 mM) for folic
acid (Goldman, 1971; Sirotnak, 1985, 1987). RFC is a member of the
solute carrier family (SLC) of facilitative carriers which currently com-
prises 300 genes that have been classified into 43 subfamilies (http://
www.gene.ucl.ac.uk/nomenclature/). The SLC family belongs to an
even larger superfamily known as the major facilitator superfamily
Reduced
folates
Folate
(oxidized and
reduced )
H
+
Folate
(mM)
H
+
Folate
(mM)
Folate
Folate
(nM)
(i.e., 5-CH
3
-THF )
Folate
(mM)
CSF Blood Intestinal lumen Upper intestinal Choroid plexus
epithelial cell
Brain
(acidic pH)
Folate
(mM)
- PCFT : proton-coupled symport of folates and protons.
- RFC : anion-exchange-based transport of reduced folates.
- FR
: high-affinity (K
D
=0.1nM)-based endocytosis of oxidized and reduced folates.
Intestinal
epithelium
(neutral pH)
Figure 4.4 Putative transport pathways of folates from the intestinal lumen through
the blood into the cerebrospinal fluid and the brain.
Adaptation to Folate Deprivation 113
(MFS) of transporters (Saier et al., 1999). Human RFC contains 591
amino acids with a core molecular weight of 64 kDa; however,
although it contains a single consensus site for N-linked glycosylation,
RFC undergoes an extensive glycosylation thereby resulting in a
molecular mass of 85 kDa (Drori et al., 2000). Putative hydropathy
plots and membrane topology studies suggested that RFC contains 12
transmembrane domains (TMDs) with a short cytosolic N-terminus and a
long cytosolic C-terminus (Ferguson and Flintoff, 1999; Liu and
Matherly, 2002). In contrast to most known folate transport systems,
RFC can neither bind nor hydrolyze ATP in order to drive folate
substrate translocation across the plasma membrane. Rather, the RFC-
dependent uphill influx of folates and antifolates is coupled to the
downhill efflux of organic phosphates including thiamine
monophosphate and pyrophosphate (Zhao et al., 2002) that are
synthesized and retained in the cytoplasm. However, when compared
Folate(glu
n
)
FPGS
GGH
Cytoplasm
ATP ADP+Pi
RFC
(AFD)
MRP1,5
BCRP
Reduced
folate
Folic acid
5-CH
3
-THF
PCFT
DHFR
Reduced
folate
Reduced
folate
FR
a
Figure 4.5 Proposed model for major adaptive responses to folate deficiency. Note
that upward arrows denote upregulation whereas downward arrows denote down-
regulation of the relevant components activity. AFD-Absolute folate deficiency.
114 Ilan Ifergan and Yehuda G. Assaraf
to transporters including MRPs, RFC is considered a low-capacity
transporter. Therefore, the net efflux of organic phosphates via the
RFC is negligible when compared to the rate of the biosynthesis of
these organic phosphates (Goldman, 1971). The central developmental
role that mammalian RFC plays has been established in RFC knockout
mice studies; whereas RFC-null embryos died in utero before embryonic
day 9.5 (E9.5), the rescue of the nullizygous embryonic lethal phenotype
was achieved by supplementation of pregnant monoallelic RFC dams
with 1 mg daily subcutaneous doses of folic acid (Zhao et al., 2001).
Furthermore, the rescued RFC nullizygous embryos died within 12
days after birth due to a failure of the hematopoietic organs (Zhao et al.,
2001).
(ii) The second route of folate uptake involves the small family of folate
receptors (FRs) (Figs. 4.4 and 4.5). This group of receptors is encoded
by three distinct genes known as FRa, FRb, and FRg, respectively
(Elnakat and Ratnam, 2004, 2006; Elwood, 1989; Lacey et al., 1989;
Ratnam et al., 1989; Sadasivan and Rothenberg, 1989; Shen et al.,
1994, 1995). An additional FR isoform (i.e., FRd) has been putatively
identified from genome database mining; however, neither its tissue
expression nor its folate-binding activity has been experimentally
established (Spiegelstein et al., 2000). Whereas the high-affinity
folate-binding membrane glycoproteins FRa and FRb are glycosyl-
phosphatidylinositol (GPI)-anchored receptors, FRg is a secreted pro-
tein which lacks a GPI anchor. FRa displays a high affinity for folic acid
and 5-CH
3
-THF (K
D
0.110 nM) but lower affinity (K
D
10
300 nM) for other reduced folates and MTX (Antony, 1992; Brigle
et al., 1994; Wang et al., 1992; Westerhof et al., 1995). The FR-
dependent uptake of folates proceeds via a classical mechanism of
receptor-mediated endocytosis. Moreover, whereas RFC exhibits a
relatively wide pattern of tissue expression, the expression of FRa,
the most abundant FR isoform in adults, is restricted to few tissues
including the apical (luminal) surface of epithelial cells. Moreover,
membrane-associated FRb is expressed and binds folate in the placenta
(Ratnam et al., 1989). A specific pathological state of loss of function of
FRa has been described recently; auto-antibodies against FRa have
been associated with cerebral folate deficiency (CFD) in the presence of
normal folate blood levels (see paragraph below discussing Cerebral folate
deficiency syndrome and auto-antibodies against folate receptors). It therefore
appears that FRa is a key mediator of folate uptake into the brain
(Ramaekers et al., 2005; Spector and Johanson, 2006). Additionally,
gene knockout studies have established that FRa is vital for normal
nerve tube development of mouse embryos; importantly, maternal folic
acid supplementation prevents the development of NTDs in FRa
knockout mouse embryos (Piedrahita et al., 1999) (see the paragraph
Adaptation to Folate Deprivation 115
below entitled Knockout mice lacking the folic acid-binding protein Folbp1,
currently termed Folr1, the murine homologue of the human FRa). In contrast
to the severe morphogenetic abnormalities of FRa knockout embryos
that die in utero, FRb knockout mouse embryos have been shown to
develop normally (Piedrahita et al., 1999).
Cerebral folate deficiency syndrome and auto-antibodies against folate receptors:
Cerebral folate deficiency (CFD) is defined as any neuropsychiatric syn-
drome associated with low levels of 5-CH
3
-THF in the cerebrospinal fluid
(CSF) (the active THF cofactor in the blood and CSF) in the presence of
normal folate metabolism outside the CNS, as evidenced by normal folate
levels in serum and erythrocytes, normal hematological values and normal
homocysteine levels. Infantile-onset CFD is a neurological syndrome that
presents 4 to 6 months after birth; the primary manifestations of this type
of CFD are significant irritability, slow head growth, cerebellar ataxia,
psychomotor retardation, pyramidal tract signs in the legs, dyskinesias, and
sometimes seizures (Ramaekers and Blau, 2004; Ramaekers et al., 2002).
Later in infancy, central visual disturbances may become apparent thereby
leading to blindness. Despite these numerous and severe clinical manifesta-
tions, the sole biochemical anomaly consistently identifiable in these CFD
patients is low levels of 5-CH
3
-THF in the CSF. Maintenance of normal
folate levels in the CSF is primarily attributable to the transport activity of
high affinity FRs that are anchored to the plasma membrane of various
epithelial and mesenchymal cells via a GPI moiety (Fig. 4.4). Membrane-
associated FRa is expressed at substantial levels in the basolateral surface of
the choroid plexus epithelium (Fig. 4.4). 5-CH
3
-THF binds to FRa with a
very high affinity (K
D
0.1 nM) and is then internalized into choroid
epithelial cells via receptor-mediated endocytosis (Fig. 4.4). As discussed
above, reduced folate transport from the plasma into the CSF is apparently a
concentrative process as the ratio of 5-CH
3
-THF levels in the CSF versus
the plasma is 3:1 (Spector, 1989; Weitman et al., 1992). Following endocy-
tosis and dissociation of 5-CH
3
-THF from FRa in the cytosol, a fraction of
the endocytotic membrane-bound FRa recycles back to the basolateral cell
surface of the choroid plexus epithelium (Holm et al., 1991); whereas,
another fraction of FRa is sorted out to the apical surface of the choroid,
where it is cleaved from its GPI anchor. Hence, receptor molecules are
released into the CSF while retaining their functional capacity to bind
5-CH
3
-THF. Then, at the apical surface of the choroid, folate efflux into
the CSF is believed to occur via the bidirectional anion exchanger RFC that
displays a folate transport K
m
at the micromolar range, consistently reflect-
ing the intracellular concentration of reduced folate cofactors (Assaraf et al.,
2006) (Fig. 4.4). Once in the spinal fluid, 5-CH
3
-THF is transported into
neuronal tissues via the RFC. Taken together, the above physiological
characteristics clearly establish the central role that FRa plays in the
116 Ilan Ifergan and Yehuda G. Assaraf
transport of 5-CH
3
-THF across the choroid plexus epithelial cells. There-
fore, this raised the hypothesis that CFD may potentially be associated with
alterations (i.e., decrease) in the capacity of FRa to transport reduced folates
across the choroid epithelium. Direct clinical evidence to support this
hypothesis was first obtained by Rothenberg and colleagues (Rothenberg
et al., 2004) who initially discovered auto-antibodies against FRs in women
with a pregnancy complicated by NTDs. In this study, it was found that 9 of
12 women with a current or previous affected pregnancy with an NTD
contained neutralizing auto-antibodies against FRs. Moreover, these auto-
antibodies were capable of blocking the binding of radiolabeled folic acid to
FRs on isolated membranes from FRa-rich cells as well as in intact KB cells
that highly express FRa. Consistently, folic acid uptake by KB cells was
potently blocked in the presence of these auto-antibodies to FRs.
Subsequent studies by Ramaekers and associates (Ramaekers et al., 2005)
revealed that sera from 25 out of 28 children with CFD contained high-
affinity blocking auto-antibodies against membrane-bound FRs present in
choroid plexus epithelial cells. Oral replacement therapy with folinic acid
improved the levels of 5-CH
3
-THF in the CSF and led to clinical amelio-
ration. Hence, these important studies demonstrated for the first time
that CFD is a CNS disorder which may be caused by auto-antibodies
abolishing the FRa-dependent transport of 5-CH
3
-THF from the plasma
into the CSF.
Knockout mice lacking the folic acid-binding protein Folbp1, currently termed
Folr1, the murine homologue of the human FRa: Early embryonic lethality was
observed in knockout mice lacking the folic acid-binding protein Folbp1
(currently termed Folr1), the murine homologue of the human FRa
(Piedrahita et al., 1999). Histological examination of transverse sections of
Folbp
/
nullizygous E8.5 embryos revealed defects in the neuroepithelium;
in the cephalic neural tube, neither the forebrain nor the optic vesicles were
formed. The neuroepithelium was only one to two cells thick. At the level
of the midbrain, the neuroepithelium of both the basal and planar plates was
also limited to no more than two cells in thickness. Remarkably, supple-
menting pregnant Folbp1
/
dams with folinic acid reversed the embryonic
lethality and the neuroepithelium defects in 80% of the nullizygous pups.
Moreover, administration to pregnant rats of an antiserum against FRs
(da Costa and Rothenberg, 1996) resulted in resorption of embryos or
multiple embryonic developmental abnormalities (da Costa et al., 2003).
Specifically, the antiserum displayed a dose-dependent effect on embryo
viability and organogenesis. Administration of folinic acid prevented tera-
togenicity resulting from low doses of antiserum, but failed to protect
against high antibody concentrations that inflicted embryo damage by
immune-mediated cell lysis. Moreover, resorption of embryos with larger
doses of antiserum was prevented by the immunosuppressant dexametha-
sone. Furthermore, cardiovascular abnormalities were observed in 100%
Adaptation to Folate Deprivation 117
of Folr1 knockout mice lacking a functional Folr1 (Zhu et al., 2007).
A dose-response study with folinic acid was also performed in order to
determine the impact of maternal folate supplementation on cardiac devel-
opment in Folr1 nullizygous fetuses. Hence, partially rescued preterm Folr1
fetuses were found to harbor outflow tract defects, aortic arch artery
abnormalities, and isolated dextrocardia (i.e., a peculiar anatomic anomaly
in which the heart is positioned on the right side of the chest). Consistent
with the above results, maternal supplementation with folinic acid rescued
embryonic lethality and the observed cardiovascular abnormalities in a
dose-dependent manner. This study further establishes the beneficial effects
of folate supplementation in the prevention of congenital heart defects. The
latter may be possibly mediated via the folate supplementation impact on
neural crest cells and regulation of genes associated with signaling pathways
resulting in normal development of pharyngeal arches and the secondary
heart field. Moreover, these studies clearly demonstrate the important
developmental role of the Folr1 in folate homeostasis.
(iii) The third route of folate uptake is the low pH folate transporter that
has been recently cloned and termed proton-coupled folate transporter
(PCFT; heme-carrier protein, HCP-1; SLC46A1) [Qiu et al., 2006);
Figs. 4.4 and 4.5]. This transport system functions optimally at acidic
pH (5.5) but poorly at physiological pH (7.4). PCFT recognizes folic
acid, reduced folates (e.g., 6S-5-CH
3
-THF) and MTX as transport
substrates with comparable high affinities (K
m
0.52 mM) (Assaraf
et al., 1998; Henderson and Strauss, 1990; Kumar et al., 1997; Sierra
et al., 1997). Hence, PCFT plays a key role in the intestinal absorption
of folates within the acidic microclimate of the upper small intestine,
notably the duodenum and the upper jejunum (Qiu et al., 2006)
(Fig. 4.4). Human PCFT is a 459 amino acids membrane glycoprotein
(50 kDa) with 12 predicted transmembrane helices. PCFT belongs
to the solute carrier superfamily (i.e., SLC) of transporters (Saier et al.,
1999). PCFT is a classic representative of proton-coupled folate trans-
porters and low pH transport systems mediating intestinal absorption of
various essential nutrients including amino acids, peptides, metal ions
and various organic anions (Boll et al., 2002; Fei et al., 1994; Gunshin
et al., 1997; Nozawa et al., 2004). Although initially deposited into the
GenBank as a heme carrier protein (HCP1) (Shayeghi et al., 2005),
further kinetic studies at pH 5.5 revealed that the predominant (if not
the sole) transport function of PCFT is high-affinity transport of folic
acid [K
m
0.83 mM; (Qiu et al., 2006)] as well as reduced folates such
as (6S)55-CH
3
-THF [K
m
0.53 mM; (Qiu et al., 2006)]. At this
acidic pH, PCFT was also shown to mediate an extremely high affinity
transport of the novel antifolate pemetrexed [Alimta; K
m
0.09 mM;
(Wang et al., 2004)] as well as high affinity for MTX [K
m
2.0 mM;
118 Ilan Ifergan and Yehuda G. Assaraf
(Qiu et al., 2006)]. PCFT displayed an optimal folic acid transport
activity at pH 5.5 and the transport V
max
markedly declined as the pH
increased toward the physiological pH of 7.4. PCFT was shown to
function as an inward symporter cotransporting protons and folates and
thus proved to be an electrogenic transporter. Hence, the downhill
gradient of protons in the acidic microclimate of the proximal small
intestine drives the uphill uptake of oxidized and reduced folates.
Consistent with the important role that PCFT plays in intestinal folate
absorption, it was recently discovered that inactivating mutations in
this folate transporter result in hereditary (congenital) folate malab-
sorption (HFM) (Qiu et al., 2006; Zhao et al., 2007) (See the detailed
paragraph below entitled Loss-of-function mutations in the PCFT gene and
HFM). Moreover, in the presence or absence of restoration of normal
folate levels in the circulation of HFM patients via oral folate replace-
ment therapy, these patients also exhibit a folate transport defect from
the blood into the CNS (Geller et al., 2002).
Loss-of-function mutations in the PCFT gene and HFM: Consistent with the
acidic pH of the upper small intestine (Mason, 1994; McEwan et al., 1990;
Selhub and Rosenberg, 1981), intestinal folate transport occurs primarily via
PCFT (Qiu et al., 2006). HFM (OMIM 229050HFM) first described by
Lubhy et al. in 1961 (Lubhy, 1961) is a rare autosomal recessive disorder
caused by impaired intestinal folate absorption and defective folate uptake
into the CNS with an early onset of a few months after birth. The disease
manifests itself with very low levels of folate in the blood as well as in the
CSF. HFM patients typically suffer from some or all of the following
symptoms: megaloblastic anemia, recurrent or chronic diarrhea, failure to
thrive, immune deficiency, recurrent infections (e.g., upper respiratory
infections), seizures, neurogical deficits as well as moderate to severe mental
retardation. The underlying deficit in this syndrome has been shown as a
severe defect in intestinal folate absorption (Geller et al., 2002) reflected in
an abnormal ratio (i.e., markedly decreased) of CSF:serum folate level.
As mentioned above, in healthy individuals, the normal folate level ratio
of CSF:serum is typically 3:1 (Alperin and Haggard, 1970), whereas in
HFM patients this ratio was found to range from 1:10 to 1:1 (Geller
et al., 2002). Importantly, if diagnosed early, most of the symptoms of this
disorder could be resolved by parenteral administration of large doses of
lecucovorin (Geller et al., 2002; Matherly and Goldman, 2003; Poncz and
Cohen, 1996). However, it should be noted that although the signs and
symptoms of this disorder could be obviated by folate replacement therapy
and the CSF:blood folate level ratio is improved, the normal CSF:blood
folate ratio of 3:1 was never restored. Recently, multiple loss-of-function
mutations were identified in the PCFT gene in five families with HFM
(Qiu et al., 2006; Zhao et al., 2007). These PCFT inactivating mutations
Adaptation to Folate Deprivation 119
were largely homozygous and were spread out throughout the entire coding
region. Introduction of these mutations into a PCFT expression vector and
their stable transfection into a HeLa subline that is doubly deficient in both
PCFT as well as RFC transport activities allowed for the assessment of the
impact of these congenital PCFT mutations on folate transport at acidic pH.
A complete loss of folic acid transport was observed in four of the five HFM
families; this loss of folate transport was due to decreased transporter stability
and/or plasma membrane trafficking defects inflicted by the amino acid
substitutions in highly conserved transporter regions. Whereas, a few
mutant PCFT transporters retained some residual folate transport activity
at low pH. Moreover, folate transport at acidic pH was markedly impaired
in immortalized lymphocytes from HFM patients. Taken together, these
pioneering studies establish the key role that inactivating PCFT mutations
play in the pathogenesis of HFM. Moreover, it is imperative that these
molecular studies will facilitate the rapid diagnosis and treatment of this
disorder in infants as well as the prenatal identification of families harboring
the mutant PCFT gene and thereby offer proper genetic counseling.
2. The role of influx transporters in adaptation to folate deficiency
Several lines of evidence support the notion that an increased activity of
folate influx systems is a crucial cellular adaptive response to folate
deficiency:
(a) The facility of RFC to mediate an efficient influx of reduced folates
should be beneficial to cells under conditions of folate deprivation
(Fig. 4.5). An increased activity of this reduced folate transporter may
serve as an important cellular adaptive response under conditions in
which folates are present in the growth medium at limiting concentra-
tions. Indeed, the T-cell leukemia subline CCRF-CEM-7A which was
gradually adapted to grow in 150-fold decreased leucovorin concen-
tration relative to its parental CCRF-CEM counterpart, displayed a
dramatic overexpression of the RFC ( Jansen et al., 1990). This over-
expression was associated with a consistently increased V
max
of MTX
influx [95-fold; (Assaraf et al., 1998)] when compared to parental
CCRF-CEM cells. Moreover, the expression of MRP1, an ATP-
dependent folate efflux transporter, was nearly completely lost in
these LF-adapted CCRF-CEM-7A cells (Assaraf et al., 2003). These
alterations were consistently associated with a marked decrease in both
folic acid and leucovorin growth requirements in CCRF-CEM-7A
cells.
(b) In independent studies, four colon cancer cell lines (C26-A, C2610,
C26-G, and WiDr) as well as three squamous cell carcinoma lines of the
head and neck (HNSCC) (11B, 14C, and 22B) were adapted to grow
in a culture medium containing LF levels (i.e., 0.52.5 nM leucovorin)
120 Ilan Ifergan and Yehuda G. Assaraf
(Backus et al., 2000). Consistent with the efficient folate influx ability of
the RFC, a two- to sevenfold increased transport activity of RFC in
LF-adapted cell lines was observed when compared to their folate-
replete counterparts. Hence, the increased activity of RFC appears to
serve as an important adaptive response to folate deficiency. The
importance of RFC transport activity in folate deplete growth condi-
tions should be studied in regards to the mechanism underlying upre-
gulation or downregulation of this folate transporter gene. Whereas
several transcript variants of the human RFC (regulated by several
promoters) have been identified, promoter B (pB) of the human
RFC was found to drive the expression of one variant which serves
as the major form of the intestinal human RFC (Gong et al., 1999;
Nguyen et al., 1997; Sirotnak and Tolner, 1999; Williams and Flintoff,
1998; Zhang et al., 1998). A recent study aimed at identifying the
minimal promoter region required for basal transcriptional activity of
this promoter undertook a deletion construct analysis of the human
RFC pB and determined their transcriptional activity in colon cancer-
derived Caco-2 cell transfectants (Subramanian et al., 2003). The mini-
mal region required for basal activity of human RFC pB in Caco-2 cells
was identified as a sequence between nucleotides 1088 and 1043.
Moreover, a sequence between nucleotides 141 and 2016 (i.e.,
outside the minimal region of the pB) was found to be responsive to
folate deficiency in Caco-2 cells; this promoter region led to a signifi-
cant and specific upregulation in folate uptake under conditions of
folate deficiency. This upregulation was associated with a parallel
increase in human RFC mRNA and protein levels as well as in the
transcriptional activity of human RFC pB. The identification of puta-
tive folate responsive elements in the human RFC pB may pave the
way for detailed studies aimed at pinpointing the exact role that
transcriptional upregulation of the human RFC plays in adaptation to
cellular conditions of folate deficiency.
(c) As with enhanced RFC transport activity, increased FR levels should
also play a contributing role in the adaptation to folate deficiency
(Fig. 4.5). This is particularly true as FRs mediate the unidirectional
influx of folates and are devoid of the folate efflux component displayed
by the RFC. Consistent with this presumption, it has been recently
shown that introduction of an FRa cDNA into cells that normally lack
receptor expression induced higher proliferation rates and increased
cellular survival under conditions of folate deprivation relative to their
parental counterparts (Antony, 1996). Indeed, a 5.5-fold increased
density of FRs was documented in human tumor xenografts implanted
in folate-deprived mice (Mendelsohn et al., 1996). Interestingly, the
increased density of FRs was associated with a modest reduction in
the affinity of this receptor to folic acid in four out of five tumors,
Adaptation to Folate Deprivation 121
suggesting an increased expression of the second FR with lower affinity
(i.e., FRb) in these tumors (Mendelsohn et al., 1996). Additionally, an
increased FPGS activity was also documented in these tumors. The
increased levels of FRs should augment the rate of unidirectional folate
uptake. Once taken up into cells, folates will be efficiently retained in
the cancer cells due to the increased activity of FPGS, the enzyme
responsible for folylpolyglutamylation (Fig. 4.5).
(d) In another study, a severe folate restriction (the growth medium
contained only 15 pM leucovorin; Lamers et al., 2006) was imposed
on Chinese hamster lung fibroblasts DC-3F/FA3 cells (Zhu et al.,
2002). Consistent with the above-mentioned experiments, these LF-
adapted cells also exhibited FRa overexpression. Moreover, three
independent LF-selected clones (i.e., FA3, FA7, and FA14) displayed
a five- to sixfold increased DHFR protein levels as well as catalytic
activity (Zhu et al., 2002). Thus, increased FRa levels should enhance
the rate of folate uptake following which the internalized folates could
be efficiently reduced and further utilized due to the increased activity
of DHFR (Fig. 4.5).
(e) In yet another study, LF [15 pM leucovorin (Lamers et al., 2006)]-
selected Chinese hamster lung fibroblasts were found to overexpress
FRa mRNA by more than 500-fold; furthermore, the mechanisms
underlying this dramatic increase in FRa transcripts were found to be
mediated via a significant gene amplification as well as an increase in
transcript half-life (Zhu et al., 2001). As has been discussed above, FR
levels have been shown to be upregulated under LF levels (Kane et al.,
1988; McHugh and Cheng, 1979). This inverse relationship was found
to occur at the translational level in cervical carcinoma cells (Sun and
Antony, 1996); specifically, it was suggested that the folate-dependent
translational regulation of FRa is attributable to an interaction of a
46 kDa cytosolic protein, recently identified as hnRNP E1(Xiao et al.,
2001), with an 18-base cis-element in the 5
0
-untranslated region of the
human FRa transcript (Sun and Antony, 1996). In contradistinction to
these results, a recent study demonstrated that 48 and 72 h of exposure
of HepG2 cells to a folate-deficient medium resulted in 12% and 43%
decreased FR protein expression, respectively (Abdel Nour et al.,
2007). Similarly, FR expression in kidney tubules was shown to be
downregulated in response to a LF diet (da Costa et al., 2000; Gates
et al., 1996). The molecular mechanisms underlying this decreased FR
protein expression under conditions of folate deficiency are yet to be
revealed.
(f ) Chinese hamster ovary (CHO) cells were exposed to a stepwise selec-
tion to the lipophilic antifolate pyrimethamine resulting in the estab-
lishment of the Pyr
R100
subline (Assaraf and Slotky, 1993). These
pyrimethamine-resistant cells displayed a fourfold increase in the folic
122 Ilan Ifergan and Yehuda G. Assaraf
acid influx mediated by the low pH folate transporter currently identi-
fied as PCFT (HCP-1/SLC46A1) (Assaraf et al., 1998; Qiu et al., 2006).
These lipophilic antifolate-resistant cells also displayed the loss of several
folate efflux transporters including MRP1 (Stark et al., 2003). The
increased activity of the folate influx transporter PCFT along with the
loss of several ATP-driven folate efflux transporters should increase
folate uptake and decrease folate efflux, respectively, thereby increasing
the intracellular folate pool (Fig. 4.5). Indeed, Pyr
R100
cells displayed a
17-fold increase in the intracellular folic acid pool relative to parental
CHO AA8 cells (Assaraf and Goldman, 1997; Stark et al., 2003).
Of special note was the finding that the expansion in the intracellular
folate pool in these cells resulted in high levels of resistance to lipophilic
antifolates via competitive negation of the antifolate inhibitory effect at
the level of folate-dependent enzymes. Hence, it appears that over-
expression of PCFT may serve as a cellular adaptive response under
folate-deprived conditions; however, further studies are warranted in
order to identify the mechanisms underlying PCFT overexpression
under conditions of folate deficiency and/or antifolate inhibition.
D. Downregulation of folate efflux systems
1. The ABC superfamily of transporters
The term ABC transporters was first introduced in 1992 by Chris Higgins in
a memorable review (Higgins, 1992). The acronym ABC represents the
highly conserved ATP-binding cassette signature of the 49 known members
of this large superfamily of human transporters (Dean and Allikmets, 2001);
the ABC is utilized by these transporters to bind and hydrolyze ATP,
thereby energizing the vectorial transport of various substrates across
biological membranes (Borst et al., 2000; Dean, 2002; Gottesman et al.,
2002). Additional terminologies have been used for this superfamily includ-
ing Traffic ATPases and P-glycoproteins (Pgps). The basic structure of the
ABC transporters is often based on minimal data; however, the putative
membrane topology of Pgp (ABCB1), one of the most famous and well-
studied members of this group of transporters, is thought to consist of 12
TMDs and two ATP-binding sites and is 1,280 amino acids long. Some
ABC transporters may dimerize to form two nearly equal (e.g., BCRP) or
unequal halves (ABCG5 and ABCG8), resulting in a homodimeric or
heterodimeric ATP-dependent transport systems, respectively. Several
ABCtransporters contain more than 12 hydrophobic transmembrane helices
and may be much larger than Pgp; for example, MRP1 is a 1522 amino acids
membrane glycoprotein with as many as 17 transmembrane helices.
ABC transporters mediate the ATP-dependent transport of various
drugs and their conjugates and thereby confer a multidrug resistance
Adaptation to Folate Deprivation 123
(MDR) phenotype upon cancer cells (Szakacs et al., 2006). This transport
activity is clearly exemplified by several MDR transporters including Pgp,
MRP1 (ABCC1), or BCRP (ABCG2/MXR/ABCP), each of which has
been already shown to mediate MDR in cancer cells. The identified genetic
variations of several ABC transporters (Fromm, 2002) may alter the
response to drug treatment. Several members of the ABC superfamily play
a key role in preventing the intestinal absorption of toxic compounds,
including many drugs and food components. Moreover, these transporters
play a key role in a second defense line by protecting vital organs in the body
including the brain, CSF, testis, as well as protecting the fetus against various
highly toxic xenobiotics. Consistent with this protective role of ABC
transporters, knockout mouse models of ABC transporter genes (e.g.,
ABCB1) resulted in an altered (i.e., decreased) bloodbrain barrier function
(Schinkel et al., 1997), intestinal drug absorption ( Jonker et al., 2002;
Sparreboom et al., 1997), fetal drug exposure (Smit et al., 1999), and
drug-induced damage to testicular tubules (Wijnholds et al., 2000). Addi-
tional physiological roles of several ABC transporters include the excretion
of endogenous metabolites from mammalian secretory epithelia, even
against a steep concentration gradient (Borst and Elferink, 2002). Accord-
ingly, ABC transporters play a key physiological role in liver excretion of
bile salts (transported by BSEP, the bile salt export pump, ABCB11),
phosphatidylcholine (MDR3 P-glycoprotein, ABCB4), bilirubin glucuro-
nides (MRP2, ABCC2), and various hydrophobic cytotoxic drugs (MDR1
P-glycoprotein, ABCB1) (Borst and Elferink, 2002). Moreover, ABC
transporters are capable of mediating hydrophobic peptide export such as
gramicidin D or cyclosporin A (transported by ABCB1) (Borgnia et al.,
1996; Sarkadi et al., 1994), peptides for antigen presentation (transported by
heterodimeric ABC transporters TAP1 and TAP2) (Schmitt and Tampe,
2000), and mitochondrial peptides (transported by ABC transporter related
to TAP) (Young et al., 2001). Nuclear receptors play a key role in the
transcriptional regulation of many mammalian ABC transporters; this is
clearly the case for MDR1 and MRP2 (Borst and Elferink, 2002).
2. The MRP (ABCC) family of MDR efflux transporters
Human MRPs constitute an important ABCC subfamily comprising 13
members including the cystic fibrosis conductance regulator (CFTR/
ABCC7) and the sulfonylurea receptors (SUR1/ABCC8 and SUR2/
ABCC9) (Deeley et al., 2006). Several MRPs mediate the ATP-driven
efflux of both endogenous compounds and various drugs as well as their
conjugated metabolites including Vinca alkaloids, anthracyclines, epipodo-
phyllotoxins, camptothecins, taxenes, antifolates, nucleoside and nucleotide
analogues, peptide-based cytotoxins, platinum compounds, and tyrosine
kinase inhibitors (Cools et al., 2005; Szakacs et al., 2006). The detoxification
process of many cytotoxic drugs requires their conjugation to glutathione
124 Ilan Ifergan and Yehuda G. Assaraf
(GSH), glucuronate, or sulfate before their binding to- and extrusion by
MRPs. However, the hydrophilic nature of the resultant conjugates pre-
vents them from diffusing back into the cell through biomembranes. Hence,
the ATP-dependent efflux of these drug conjugates must be mediated by
dedicated transporters, as first pointed out by Ishikawa (1992). As discussed
above, various members of the MRP family mediate the export of these
drug conjugates. Furthermore, the broad spectrum of MRP substrates also
includes phosphate as well as glutamate conjugates of organic compounds
(Borst et al., 2004; Keppler et al., 1997; Kruh and Belinsky, 2003; Wielinga
et al., 2005). MRP1 (ABCC1) was the first GS-X pump to be identified in
cells selected for MDR (Cole et al., 1992). This 190 kDa N-glycosylated
transmembrane protein is the founding member of the MRP family, first
cloned in the laboratory of Cole et al. (1992). Transport experiments carried
out with purified membrane vesicles from MRP1-overexpressing cells
established that MRP1 mediates the transport of various drugs conjugated
to GSH, sulfate, or glucuronate. MRP2 (ABCC2) was cloned in 1996
(Konig et al., 1996; Paulusma et al., 1996) and MRP3, MRP4, and
MRP5 soon followed (Kool et al., 1997) due to the identification of 21
potential human ABC transporters by Allikmets et al. (1996). Additional
studies identified new MRPs including MRP6, MRP7 (Hopper et al.,
2001), MRP8, and MRP9, respectively (Tammur et al., 2001). MRP4
(ABCC4) and MRP5 (ABCC5) were found to mediate the transport of
cyclic nucleotides and nucleotide analogs and thus confer resistance to base-
and nucleoside analogues used in the chemotherapy of cancer and viral
diseases such as HIV and AIDS. Whereas these 9 MRPs have been identi-
fied experimentally, an additional MRP pseudogene has been identified
during a computerized human genome search, suggesting that all human
MRPs have been already identified.
The MRP family is composed of two major structural types, one with 17
hydrophobic transmembrane segments (MRP1, -2, -3, -6, and -7), and one
with 12 transmembrane segments (MRP4, 5, -8, -9, and -10). Moreover,
the membrane-spanning domains (MSDs) of various MDR transporters
play a key role in substrate binding as was revealed in studies using affinity
probes as well as site-directed mutagenesis (Hafkemeyer et al., 1998; Loo
and Clarke, 1994).
3. BCRP (ABCG2), a MDR efflux transporter
Breast cancer resistance protein (BCRP) is a member of the small ABCG
subfamily of transporters which is composed of five members including
ABCG1, ABCG2, ABCG4, ABCG5, and ABCG8, respectively. Whereas
Pgp and MRP1 confer drug resistance upon most cell lines, the mechanism
underlying high level mitoxantrone-resistance and lower level resistance to
anthracyclines and camptothecins could not be attributed to these transpor-
ters (Borst and Elferink, 2002). This initially mysterious and atypical
Adaptation to Folate Deprivation 125
drug resistance phenotype was found to be a result of decreased drug
accumulation which could be explained by transport activity of an uniden-
tified efflux transporter. Finally, a 72 kDa membrane glycoprotein was
cloned by Doyle et al. (1998) and termed BCRP. The latter is a half-size
ABC transporter which is expressed at substantial levels in breast cancer
cells. Allikmets et al. (1998) termed this transporter ABCP, a transporter
present at high levels in the placenta. BCRP cDNA was subsequently
cloned from human placenta (Allikmets et al., 1998) and mitoxantrone-
resistant human colon carcinoma cells (Miyake et al., 1999). The BCRP
gene is composed of 16 exons and 15 introns spanning 66 kb and has been
mapped to chromosome 4q22 (Bailey-Dell et al., 2001). Initial studies of the
BCRP promoter region revealed that it is TATA-less with five putative Sp1
sites downstream from a CpG island and contains several AP1 sites as well
as a functional estrogen response element (ERE) (Bailey-Dell et al., 2001;
Ee et al., 2004).
Members of the ABCG subfamily are considered half-transporters
with a single hydrophobic MSD and two cytosolic nucleotide-binding
folds (NBF; i.e., ATP-binding sites). It has been found that the transport
activity of these transporters is activated upon dimerization (Polgar et al.,
2004, 2006). BCRP transports a wide array of both positively and nega-
tively charged compounds including various chemotherapeutic cytotoxic
agents. Similarly to Pgp, BCRP does not require GSH for translocation of
electroneutral amphipathic drugs (Maliepaard et al., 2001). The broad
spectrum of BCRP substrates includes anticancer drugs such as anthracy-
clines (e.g., mitoxantrone and doxorubicin), epipodophyllotoxins, irinote-
can (SN-38) and topotecan, bisantrene, imatinib, flavopiridol, hydrophilic
antifolates (e.g., MTX and pemetrexed) and lipophilic antifolates (e.g.,
trimetrexate and pyrimethamine) and their di- and triglutamate conjugates
as well as the nucleoside analogues lamivudine and zidovudine as has been
previously reviewed (Krishnamurthy and Schuetz, 2006). Additionally,
BCRP exports various dietary toxic compounds such as the carcinogen
PhIP, the chlorophyll derivative pheophorbide as well as porphyrins.
BCRP also extrudes fluorescent compounds including Hoechst 33342,
lysotracker, rhodamine 123, and novel receptor-targeted agents such
as imatinib and gefitinib (Morgillo and Lee, 2005; Ozvegy et al., 2001).
A recent study demonstrated that BCRP is expressed in mammary gland
alveolar epithelial cells during pregnancy and lactation ( Jonker et al., 2005).
In contrast to the detoxifying role of BCRP in adult humans ( Jonker et al.,
2002; van Herwaarden et al., 2003), BCRP expressed in the mammary
gland was found to secrete a variety of drugs, toxins, and carcinogens into
milk and thus expose suckling newborn and young to xenotoxins ( Jonker
et al., 2005). The contradiction between the detoxifying and the newborn
contaminating role of BCRP in humans has been partially resolved by an
additional finding from this group demonstrating that BCRP secretes
126 Ilan Ifergan and Yehuda G. Assaraf
riboflavin (vitamin B2) into milk, thereby supplying the suckling infants and
young with this important micronutrient (van Herwaarden et al., 2007).
4. The role of efflux transporters in folate homeostasis
and adaptation to folate deficiency
The intracellular folate pool can be modulated by the above-mentioned
folate influx systems, by FPGS activity as well as by several ATP-driven
folate efflux transporters of the ABC superfamily (Borst and Elferink, 2002).
Various studies demonstrated that inverted membrane vesicles isolated from
cell lines overexpressing MRP1 (ABCC1) through MRP5 (ABCC5) have
the capacity to accumulate folate and MTX in an ATP-dependent manner
(Chen et al., 2002; Hooijberg et al., 1999; Kool et al., 1999; Wielinga et al.,
2005; Zeng et al., 1999, 2001). Moreover, kinetic analyses with inside-out
membrane vesicles isolated from MRP3- and MRP5-overexpressing cells
revealed that these MRPs mediate the transport of both folic acid and
leucovorin with K
m
values in the millimolar range (K
m
0.62 mM)
(Zeng et al., 2001). cMOAT (MRP2) has been found to mediate the
transport of various reduced folate cofactors including THF, 5-CH
3
-THF,
5,10-CH
2
-THF, and 5-CHO-THF (Kusuhara et al., 1998). Kinetic studies
with inside-out membrane vesicles isolated from MRP4-overexpressing
cells documented a higher affinity for folic acid (K
m
0.17 mM) and
leucovorin (K
m
0.64 mM) than MRP3 and MRP5 (Chen et al., 2002).
The higher affinity of MRP4 for these folate derivatives is consistent with
the relatively high affinity of this efflux transporter for MTX (K
m
0.22 mM) (Chen et al., 2002). Moreover, folic acid transport by MRP3,
MRP4, and MRP5 occurs with a relatively high V
max
of 0.71.7 nmol/mg
protein/min. Thus, the ATP-dependent transport of folic acid and leucov-
orin via MRP3, MRP4, and, MRP5 occurs with a low affinity, yet high
capacity.
Kinetic studies with inside-out vesicles isolated from BCRP-overex-
pressing HEK293/ABCG2 cells documented an ATP-dependent transport
of radiolabeled folic acid with a transport rate of 87 pmol/mg protein/min
(Chen et al., 2003). In contrast to the wild-type BCRP, inside-out vesicles
isolated from HEK293/G482 ABCG2 cells overexpressing the mutant
Gly482 BCRP were devoid of folic acid transport capability. The activity
of folate efflux systems may decrease the intracellular folate pool; indeed,
overexpression of MRP1, MRP2, and MRP3 in human ovarian carcinoma
2008 cells resulted in a 3238% decrease in the intracellular folate pools
under folate-replete growth conditions (Hooijberg et al., 2003). Consis-
tently, a marked increase in the folic acid (4 h pulse) growth requirement
was documented for MRP1- and MRP3-overexpressing cells relative
to parental 2008 ovarian carcinoma cells. These results suggest that
MRP1, MRP3 as well as additional folate efflux transporters may decrease
intracellular folate pools and thereby exert a deleterious effect on cells,
Adaptation to Folate Deprivation 127
particularly under conditions of folate deficiency. Several studies lend
support to the role of various MRPs as well as BCRP in adaptation to
folate deficiency:
(a) The capacity of MRP1 to transport folates should be deleterious to cells
under conditions of folate deprivation. This possible contribution of
MRP1 to cellular folate homeostasis could be readily examined in
CCRF-CEM leukemia T-cells lacking MRP2, MRP3, and BCRP
expression but barely expressing MRP4 and MRP5 (Assaraf et al.,
2003). Accordingly, gradual folate deprivation of this cell line resulted
in the CCRF-CEM-7A cell line which can grow in about 150-fold
decreased leucovorin concentration (i.e., 0.25 nM) ( Jansen et al., 1990);
these cells showed dramatic decreases in both folic acid and leucovorin
growth requirements. As expected, MRP1 expression was nearly
completely lost (Assaraf et al., 2003) along with a dramatic overexpres-
sion of the RFC ( Jansen et al., 1990). However, the poorly expressed
transporters MRP4 and MRP5 retained their low expression level in
this LF-adapted cell line. Consistently, the loss of MRP1 expression was
associated with diminished folate efflux capability in CCRF-CEM-7A
cells; these cells showed a fivefold fall in the folic acid efflux rate
constant relative to parental CCRF-CEM cells. A similar fall in the
folic acid efflux rate was achieved by incubating parental CCRF-CEM
cells with probenecid, an MRP1 transport inhibitor (Hooijberg et al.,
1999). These results suggest that under conditions of folate deprivation,
loss of MRP1 expression serves as an adaptive response aimed at pre-
venting the deleterious folate depletion effect of this ATP-driven folate
efflux transporter (Fig. 4.5).
(b) The loss of MRP1 protein levels under folate-deprived conditions
raised the possibility that a folate-dependent bidirectional regulatory
mechanism controlling MRP1 expression is operative in these cells.
Therefore, folate-replete conditions (5 nM leucovorin) were imposed
on CCRF-CEM-7A cells which, as shown above, have completely lost
MRP1 expression under folate deplete conditions. Indeed, under these
folate-replete conditions, CCRF-CEM-7A cells displayed a complete
restoration of parental MRP1 levels (Assaraf et al., 2003). Similarly,
RFC-defective CCRF-CEM sublines (CEM/MTX-LF and CEM/
GW-70LF) lost MRP1 expression under low folic acid conditions
(25 nM folic acid). Consistently, full restoration of MRP1 expression
was achieved upon growth of these cell lines under folate-replete
conditions (2.3 mM folic acid; Assaraf et al., 2003). These compelling
evidences suggest that changes in extracellular folate status can modu-
late MRP1 levels. The molecular mechanism underlying this folate-
dependent bidirectional regulatory mechanism of MRP1 expression is
yet to be revealed.
128 Ilan Ifergan and Yehuda G. Assaraf
(c) An increased intracellular folate pool serves as a resistance mechanism
against lipophilic antifolates such as pyrimethamine and trimetrexate by
competitively negating the inhibitory effect of these antifolates. This
study supports the putative intracellular folate depletion effect of MRP1
and MRP5. Specifically, CHO cells were exposed to a stepwise selec-
tion with the lipophilic antifolate pyrimethamine, resulting in the
establishment of the Pyr
R100
subline (Assaraf and Slotky, 1993). These
pyrimethamine-resistant cells displayed a complete loss of MRP1
expression and a marked decrease in MRP5 levels (Stark et al., 2003)
along with a fourfold increase in PCFT folate influx activity (Assaraf
et al., 1998; Qiu et al., 2006). The lack of MRP1 and MRP5 expression
along with the poor expression of other MRPs, was associated with a
fivefold decreased efflux of both folic acid and cholic acid as well as
17-fold increase in the intracellular pool of folic acid relative to parental
CHO cells (Assaraf and Goldman, 1997; Stark et al., 2003). Consis-
tently, this resulted in a 14-fold decreased folic acid and leucovorin
growth requirements; moreover, a very low concentration of 100 pM
leucovorin was sufficient to support the growth of Pyr
R100
cells. These
results strongly suggest that downregulation of MRP1 and MRP5 may
be crucial components of cellular adaptation to folate deficiency.
(d) Despite the above body of evidence, folate deprivation does not neces-
sarily result in the loss of MRP1 expression as has been found with
MRP1-overexpressing 2008/MRP1 cells subjected to a pulse exposure
in folate-free medium(Hooijberg et al., 2004). Consistent with previous
studies, these folate-deprived cells showed a decreased efflux of dauno-
rubicin, a typical reflection of decreased MRP1 activity (Hooijberg
et al., 2004). Moreover, full restoration of MRP1-dependent daunoru-
bicin efflux activity was achieved upon a 48 h pulse replenishment of
folate-deprived cells with 2.5 mM leucovorin. However, this folate-
dependent MRP1 activity occurred in the absence of any alterations
in MRP1 protein levels. This folate-dependent modulation of MRP1
activity may be possibly explained by putative posttranslational modifi-
cations of MRP1 (Hooijberg et al., 2004). Indeed, a cyclic phosphory-
lation (and dephosphorylation) of MRP1 has been previously suggested as
a posttranslational modification that could modulate MRP1-dependent
anthracycline efflux activity (Ma et al., 1995). Accordingly, these
researchers showed that treatment of MRP1-overexpressing HL60/
ADR cells with the protein serine kinase inhibitors H-7 and stauros-
porine completely abolished MRP1 serine phosphorylation; in turn, the
lack of MRP1 serine phosphorylation was associated with a major
decrease in MRP1 drug efflux activity and a marked increase in drug
accumulation. The suggestion that posttranslational modifications of
MRP1 could modulate its transport activity is consistent with studies
demonstrating that blocking phosphoinositol-3-kinase (PI3K/Akt)
Adaptation to Folate Deprivation 129
activity with LY294002 in MRP1-overexpressing colon carcinoma
cells resulted in a marked increase in doxorubicin accumulation and
cytotoxicity (Abdul-Ghani et al., 2006). Yet, further studies are neces-
sary to determine whether phosphorylation (and depohosphorylation)
of MRP1 and/or other posttranslational modifications can modulate
MRP1 transport activity in a bidirectional manner thereby serving as a
delicate and well-controlled folate homeostasis mechanism.
(e) The role of BCRP in folate homeostasis was studied as well; in these
studies, MCF-7 breast cancer cells, with low BCRP protein levels, and
their MR (i.e., mitoxantrone, an anticancer drug which is a specific
BCRP substrate)-resistant MCF-7/MR subline, with BCRP overex-
pression, were gradually deprived (over 3.5 months) of folic acid from
2.3 mM to 3 nM resulting in the sublines MCF-7/LF and MCF-7/
MR-LF, respectively. These LF-adapted sublines displayed only resid-
ual BCRP mRNA consistent with poor BCRP protein levels and
transport activity along with a significant increase in FPGS activity
(Ifergan et al., 2004). The loss of BCRP and MRP1 expression along
with the increased FPGS activity were associated with enhanced cellu-
lar accumulation of folates. These results suggest that significant expres-
sion of BCRP and MRP1 may induce intracellular folate depletion.
Thus, BCRP and MRP1 are likely to undergo downregulation upon
folate deprivation (Fig. 4.5). Additional support to this hypothesis was
gained by conducting experiments of short-term (2 weeks) exposure of
MCF-7/MR cells to folate-free medium followed by one week of
adaptation to low folic acid (1 nM)-containing medium (Ifergan et al.,
2005). The short-term folate-deprived cells were characterized by a
selective confinement of BCRP to the endoplasmic reticulum instead
of the plasma membrane as was apparent to some extent before folate
deprivation. Moreover, consistent with the effects of long-term folate
deprivation, these cells also displayed a threefold decrease in BCRP and
MRP1 protein levels, thereby resulting in a 5-fold increased cellular
accumulation of folic acid. These data suggest that lack of plasma mem-
brane targeting of BCRP may be selected or regulated upon short-term
folic acid deprivation due to the folate-depleting effect of this efflux
transporter. One possible mechanismunderlying translocation of BCRP
from the plasma membrane to the cytoplasmic compartment may be
associated with downregulation of the PI3K-Akt signaling pathway as
has been documented in hematopoietic stem cells (Mogi et al., 2003).
However, further studies are required to determine whether the con-
finement of BCRP to the cytoplasmic compartment in the short-term
folate-deprived breast cancer cells may be a result of decreased activity of
the PI3K-Akt signaling pathway. Hence, consistent with the folate
transport capability of both MRP1 and BCRP, downregulation of
these ATP-driven folate efflux transporters, increased FPGS activity
130 Ilan Ifergan and Yehuda G. Assaraf
and selective confinement of BCRP to the endoplasmic reticulum
appear to be crucial components of the preservation of the precious
cellular folate pools that are particularly shrunken under conditions of
folate deficiency (Assaraf et al., 2003; Liani et al., 2003; Rothem et al.,
2002). Moreover, it has been reported recently that the folic acid trans-
port rate via BCRP is 2- to 5-fold higher at pH 5.5 than at pH 7.3
(Breedveld et al., 2007). Hence, it is possible that the role of BCRP in
folate homeostasis is mostly important in many tumor tissues, where
the extracellular pH is more acidic than in most normal mammalian
cells and may be as low as pH 5.8 (Tannock and Rotin, 1989) due to the
high rate of lactic acid production, even under aerobic conditions
(Boyer and Tannock, 1992; Gatenby and Gillies, 2004; Tannock and
Rotin, 1989).
(f ) As has been previously discussed, RFCis a nonconcentrative, facilitative
bidirectional anion-exchanger that equally displays a high-affinity (i.e.,
at the micromolar range) influx and efflux activities for reduced folate
cofactors. Hence, under conditions of extreme folate deprivation, RFC
activity may be downregulated in order to protect cells from further
loss of intracellular folates (Fig. 4.5). Indeed, we have recently found
(I. Ifergan and Y. G. Assaraf, unpublished data) a 15-fold decrease in
the viable cell number of RFC-overexpressing CHO C5/RFC cells
when compared to the RFC-deficient C5 cells after 6 days of in-
cubation in folic acid-free medium. Moreover, both the moderately
RFC-expressing as well as the RFC-overexpressing MCF7/MR and
CCRF-CEM/7A cell lines, respectively, showed 2.5-fold decreased
RFC mRNA levels after 7 and 3 days of incubation in folic acid-free
medium, respectively. These decreased mRNA levels were associated
with a marked decrease in RFCtransport activity. Hence, it appears that
whereas increased activity of RFC may serve as an important mecha-
nism of adaptation to conditions of folate deprivation ( Jansen et al.,
1990), the folate efflux component of the RFC may be detrimental to
cells under the complete lack of folates in the growth medium. Hence,
downregulation of RFC transport activity may serve as a novel
component of adaptation to conditions of severe folate deficiency.
REFERENCES
Abdel Nour, A. M., Ringot, D., Gueant, J. L., and Chango, A. (2007). Folate receptor and
human reduced folate carrier expression in HepG2 cell line exposed to fumonisin B1 and
folate deficiency. Carcinogenesis 28, 22912297.
Abdul-Ghani, R., Serra, V., Gyorffy, B., Jurchott, K., Solf, A., Dietel, M., and Schafer, R.
(2006). The PI3K inhibitor LY294002 blocks drug export from resistant colon carcinoma
cells overexpressing MRP1. Oncogene 25(12), 17431752.
Adaptation to Folate Deprivation 131
Adams, J. F., Tankel, H. I., and MacEwan, F. (1970). Estimation of the total body vitamin
B12 in the live subject. Clin. Sci. 39(1), 107113.
Allegra, C. J., Chabner, B. A., Drake, J. C., Lutz, R., Rodbard, D., and Jolivet, J. (1985).
Enhanced inhibition of thymidylate synthase by methotrexate polyglutamates. J. Biol.
Chem. 260(17), 97209726.
Allegra, C. J., Hoang, K., Yeh, G. C., Drake, J. C., and Baram, J. (1987). Evidence for direct
inhibition of de novo purine synthesis in human MCF-7 breast cells as a principal mode of
metabolic inhibition by methotrexate. J. Biol. Chem. 262(28), 1352013526.
Alperin, J. B., and Haggard, M. E. (1970). Cerebrospinal fluid folate (CFA) and the blood-
brain barrier. Clin. Res. 40, 1823.
Antony, A. C. (1992). The biological chemistry of folate receptors. Blood 79(11),
28072820.
Antony, A. C. (1996). Folate receptors. Annu. Rev. Nutr. 16, 501521.
Arnesen, E., Refsum, H., Bonaa, K. H., Ueland, P. M., Forde, O. H., and Nordrehaug, J. E.
(1995). Serum total homocysteine and coronary heart disease. Int. J. Epidemiol. 24(4),
704709.
Assantachai, P., and Lekhakula, S. (2007). Epidemiological survey of vitamin deficiencies in
older Thai adults: implications for national policy planning. Public Health Nutr. 10(1),
6570.
Assaraf, Y. G. (2006). The role of multidrug resistance efflux transporters in antifolate
resistance and folate homeostasis. Drug Resist. Updat. 9(4-5), 227246.
Assaraf, Y. G. (2007). Molecular basis of antifolate resistance. Cancer Metastasis. Rev. 26(1),
153181.
Assaraf, Y. G., Babani, S., and Goldman, I. D. (1998). Increased activity of a novel low pH
folate transporter associated with lipophilic antifolate resistance in chinese hamster ovary
cells. J. Biol. Chem. 273(14), 81068111.
Assaraf, Y. G., and Goldman, I. D. (1997). Loss of folic acid exporter function with markedly
augmented folate accumulation in lipophilic antifolate-resistant mammalian cells. J. Biol.
Chem. 272(28), 1746017466.
Assaraf, Y. G., Ifergan, I., Kadry, W. N., and Pinter, R. Y. (2006). Computer modelling of
antifolate inhibition of folate metabolism using hybrid functional petri nets. J. Theor. Biol.
240(4), 637647.
Assaraf, Y. G., Rothem, L., Hooijberg, J. H., Stark, M., Ifergan, I., Kathmann, I.,
Dijkmans, B. A., Peters, G. J., and Jansen, G. (2003). Loss of multidrug resistance protein
1 expression and folate efflux activity results in a highly concentrative folate transport in
human leukemia cells. J. Biol. Chem. 278(9), 66806686.
Assaraf, Y. G., and Slotky, J. I. (1993). Characterization of a lipophilic antifolate resistance
provoked by treatment of mammalian cells with the antiparasitic agent pyrimethamine.
J. Biol. Chem. 268(6), 45564566.
B., S. (1995). Folate chemistry and metabolism. In Folate in health and disease.
(L. B. Bailey, ed.), pp. 122. New York, Marcel Dekker.
Backus, H. H., Pinedo, H. M., Wouters, D., Padron, J. M., Molders, N., van Der
Wilt, C. L., van Groeningen, C. J., Jansen, G., and Peters, G. J. (2000). Folate depletion
increases sensitivity of solid tumor cell lines to 5-fluorouracil and antifolates. Int. J. Cancer
87(6), 771778.
Bailey-Dell, K. J., Hassel, B., Doyle, L. A., and Ross, D. D. (2001). Promoter characteriza-
tion and genomic organization of the human breast cancer resistance protein (ATP-
binding cassette transporter G2) gene. Biochim. Biophys. Acta 1520(3), 234241.
Bailey, L. B., Wagner, P. A., Christakis, G. J., Davis, C. G., Appledorf, H., Araujo, P. E.,
Dorsey, E., and Dinning, J. S. (1982). Folacin and iron status and hematological findings
in black and Spanish-American adolescents from urban low-income households.
Am. J. Clin. Nutr. 35(5), 10231032.
132 Ilan Ifergan and Yehuda G. Assaraf
Bertino, J. R. (1993). Karnofsky memorial lecture. Ode to methotrexate. J. Clin. Oncol.
11(1), 514.
Birn, H. (2006). The kidney in vitamin B12 and folate homeostasis: characterization of
receptors for tubular uptake of vitamins and carrier proteins. Am. J. Physiol. Renal.
Physiol. 291(1), F22F36.
Blom, H. J., Shaw, G. M., den Heijer, M., and Finnell, R. H. (2006). Neural tube defects
and folate: Case far from closed. Nat. Rev. Neurosci. 7(9), 724731.
Boll, M., Foltz, M., Rubio-Aliaga, I., Kottra, G., and Daniel, H. (2002). Functional
characterization of two novel mammalian electrogenic proton-dependent amino acid
cotransporters. J. Biol. Chem. 277(25), 2296622973.
Borgnia, M. J., Eytan, G. D., and Assaraf, Y. G. (1996). Competition of hydrophobic
peptides, cytotoxic drugs, and chemosensitizers on a common P-glycoprotein pharma-
cophore as revealed by its ATPase activity. J. Biol. Chem. 271(6), 31633171.
Borst, P., Balzarini, J., Ono, N., Reid, G., de Vries, H., Wielinga, P., Wijnholds, J., and
Zelcer, N. (2004). The potential impact of drug transporters on nucleoside-analog-based
antiviral chemotherapy. Antiviral Res. 62(1), 17.
Borst, P., and Elferink, R. O. (2002). Mammalian ABC transporters in health and disease.
Annu. Rev. Biochem. 71, 537592.
Borst, P., Evers, R., Kool, M., and Wijnholds, J. (2000). A family of drug transporters: The
multidrug resistance-associated proteins. J. Natl. Cancer Inst. 92(16), 12951302.
Boyer, M. J., and Tannock, I. F. (1992). Regulation of intracellular pH in tumor cell lines:
Influence of microenvironmental conditions. Cancer Res. 52(16), 44414447.
Breedveld, P., Pluim, D., Cipriani, G., Dahlhaus, F., van Eijndhoven, M. A., de Wolf, C. J.,
Kuil, A., Beijnen, J. H., Scheffer, G. L., Jansen, G., Borst, P., and Schellens, J. H. (2007).
The effect of low pH on breast cancer resistance protein (ABCG2)-mediated transport of
methotrexate, 7-hydroxymethotrexate, methotrexate diglutamate, folic acid, mitoxan-
trone, topotecan, and resveratrol in in vitro drug transport models. Mol. Pharmacol. 71(1),
240249.
Brigle, K. E., Spinella, M. J., Westin, E. H., and Goldman, I. D. (1994). Increased expression
and characterization of two distinct folate binding proteins in murine erythroleukemia
cells. Biochem. Pharmacol. 47(2), 337345.
Carmel, R., Green, R., Rosenblatt, D. S., and Watkins, D. (2003). Update on cobalamin,
folate, and homocysteine. Hematology Am. Soc. Hematol. Educ. Program 6281.
Chandler, C. J., Wang, T. T., and Halsted, C. H. (1986). Pteroylpolyglutamate hydrolase
from human jejunal brush borders. Purification and characterization. J. Biol. Chem. 261
(2), 928933.
Chen, Z. S., Lee, K., Walther, S., Raftogianis, R. B., Kuwano, M., Zeng, H., and
Kruh, G. D. (2002). Analysis of methotrexate and folate transport by multidrug resistance
protein 4 (ABCC4): MRP4 is a component of the methotrexate efflux system. Cancer
Res. 62(11), 31443150.
Chen, Z. S., Robey, R. W., Belinsky, M. G., Shchaveleva, I., Ren, X. Q., Sugimoto, Y.,
Ross, D. D., Bates, S. E., and Kruh, G. D. (2003). Transport of methotrexate, methotrex-
ate polyglutamates, and 17beta-estradiol 17-(beta-D-glucuronide) by ABCG2: Effects of
acquired mutations at R482 on methotrexate transport. Cancer Res. 63(14), 40484054.
Cole, S. P., Bhardwaj, G., Gerlach, J. H., Mackie, J. E., Grant, C. E., Almquist, K. C.,
Stewart, A. J., Kurz, E. U., Duncan, A. M., and Deeley, R. G. (1992). Overexpression of
a transporter gene in a multidrug-resistant human lung cancer cell line. Science 258(5088),
16501654.
Cools, J., Maertens, C., and Marynen, P. (2005). Resistance to tyrosine kinase inhibitors:
Calling on extra forces. Drug Resist. Updat. 8(3), 119129.
da Costa, M., and Rothenberg, S. P. (1996). Purification and characterization of folate
binding proteins from rat placenta. Biochim. Biophys. Acta 1292(1), 2330.
Adaptation to Folate Deprivation 133
da Costa, M., Rothenberg, S. P., Sadasivan, E., Regec, A., and Qian, L. (2000). Folate
deficiency reduces the GPI-anchored folate-binding protein in rat renal tubules. Am. J.
Physiol. Cell Physiol. 278(4), C812C821.
da Costa, M., Sequeira, J. M., Rothenberg, S. P., and Weedon, J. (2003). Antibodies to
folate receptors impair embryogenesis and fetal development in the rat. Birth Defects Res.
A Clin. Mol. Teratol. 67(10), 837847.
Dean, M. (2002). The human ATP-binding cassette (ABC) transporter superfamily NCBI
Monographs. National Library of Medicine, Bethesda, MD, USA.
Dean, M., and Allikmets, R. (2001). Complete characterization of the human ABC gene
family. J. Bioenerg. Biomembr. 33(6), 475479.
Deeley, R. G., Westlake, C., and Cole, S. P. (2006). Transmembrane transport of endo- and
xenobiotics by mammalian ATP-binding cassette multidrug resistance proteins. Physiol.
Rev. 86(3), 849899.
Dembo M, S. F. (1984). Membrane transport of folate compounds in mammalian cells.
Dembo M, S. F. (1984). Membrane transport of folate compounds in mammalian cells.
Folate Antagonists as Therapeutic Agents. Vol. 1. New York, Academic.
Doyle, L. A., Yang, W., Abruzzo, L. V., Krogmann, T., Gao, Y., Rishi, A. K., and
Ross, D. D. (1998). A multidrug resistance transporter from human MCF-7 breast cancer
cells. Proc. Natl. Acad. Sci. USA 95(26), 1566515670.
Drori, S., Sprecher, H., Shemer, G., Jansen, G., Goldman, I. D., and Assaraf, Y. G. (2000).
Characterization of a human alternatively spliced truncated reduced folate carrier increas-
ing folate accumulation in parental leukemia cells. Eur. J. Biochem. 267(3), 690702.
Durand, P., Prost, M., Loreau, N., Lussier-Cacan, S., and Blache, D. (2001). Impaired
homocysteine metabolism and atherothrombotic disease. Lab. Invest. 81(5), 645672.
Ee, P. L., Kamalakaran, S., Tonetti, D., He, X., Ross, D. D., and Beck, W. T. (2004).
Identification of a novel estrogen response element in the breast cancer resistance protein
(ABCG2) gene. Cancer Res. 64(4), 12471251.
Elnakat, H., and Ratnam, M. (2004). Distribution, functionality and gene regulation of folate
receptor isoforms: Implications intargetedtherapy. Adv. Drug. Deliv. Rev. 56(8), 10671084.
Elnakat, H., and Ratnam, M. (2006). Role of folate receptor genes in reproduction and
related cancers. Front Biosci. 11, 506519.
Elwood, P. C. (1989). Molecular cloning and characterization of the human folate-binding
protein cDNA from placenta and malignant tissue culture (KB) cells. J. Biol. Chem. 264
(25), 1489314901.
Fei, Y. J., Kanai, Y., Nussberger, S., Ganapathy, V., Leibach, F. H., Romero, M. F.,
Singh, S. K., Boron, W. F., and Hediger, M. A. (1994). Expression cloning of a
mammalian proton-coupled oligopeptide transporter. Nature 368(6471), 563566.
Ferguson, P. L., and Flintoff, W. F. (1999). Topological and functional analysis of the human
reduced folate carrier by hemagglutinin epitope insertion. J. Biol. Chem. 274(23),
1626916278.
Fromm, M. F. (2002). The influence of MDR1 polymorphisms on P-glycoprotein expres-
sion and function in humans. Adv. Drug. Deliv. Rev. 54(10), 129512310.
Gatenby, R. A., and Gillies, R. J. (2004). Why do cancers have high aerobic glycolysis?
Nat. Rev. Cancer 4(11), 891899.
Gates, S. B., Mendelsohn, L. G., Shackelford, K. A., Habeck, L. L., Kursar, J. D.,
Gossett, L. S., Worzalla, J. F., Shih, C., and Grindey, G. B. (1996). Characterization of
folate receptor from normal and neoplastic murine tissue: Influence of dietary folate on
folate receptor expression. Clin. Cancer Res. 2(7), 11351141.
Geller, J., Kronn, D., Jayabose, S., and Sandoval, C. (2002). Hereditary folate malabsorption:
Family report and review of the literature. Medicine (Baltimore) 81(1), 5168.
Goldman, I. D. (1971). The characteristics of the membrane transport of amethopterin and
the naturally occurring folates. Ann. NY Acad. Sci. 186, 400422.
134 Ilan Ifergan and Yehuda G. Assaraf
Goldman, I. D., and Matherly, L. H. (1986). The cellular pharmacology of methotrexate Int.
Encyl. Pharmacol. Ther., Vol. 118. Oxford/NewYork, Pergamon. pp 283302.
Gong, M., Cowan, K. H., Gudas, J., and Moscow, J. A. (1999). Isolation and characteriza-
tion of genomic sequences involved in the regulation of the human reduced folate carrier
gene (RFC1). Gene. 233(1-2), 2131.
Gottesman, M. M., Fojo, T., and Bates, S. E. (2002). Multidrug resistance in cancer: Role of
ATP-dependent transporters. Nat. Rev. Cancer 2(1), 4858.
Graham, I. M., Daly, L. E., Refsum, H. M., Robinson, K., Brattstrom, L. E., Ueland, P. M.,
Palma-Reis, R. J., Boers, G. H. J., Sheahan, R. G., Israelsson, B., Uiterwaal, C. S.,
Meleady, R., et al. (1997). Plasma homocysteine as a risk factor for vascular disease. The
European Concerted Action Project. JAMA 277, 17751781.
Graham, I. M., and OCallaghan, P. (2000). The role of folic acid in the prevention of
cardiovascular disease. Curr. Opin. Lipidol. 11(6), 577587.
Group, V. S. R. (1991). Prevention of neural tube defects: results of the Medical Research
Council Vitamin Study. MRC Lancet Jul 20; 338(8760), 131137.
Gunshin, H., Mackenzie, B., Berger, U. V., Gunshin, Y., Romero, M. F., Boron, W. F.,
Nussberger, S., Gollan, J. L., and Hediger, M. A. (1997). Cloning and characterization of
a mammalian proton-coupled metal-ion transporter. Nature 388(6641), 482488.
Hafkemeyer, P., Dey, S., Ambudkar, S. V., Hrycyna, C. A., Pastan, I., and
Gottesman, M. M. (1998). Contribution to substrate specificity and transport of non-
conserved residues in transmembrane domain 12 of human P-glycoprotein. Biochemistry
37(46), 1640016409.
Hayashi, I., Sohn, K. J., Stempak, J. M., Croxford, R., and Kim, Y. I. (2007). Folate
deficiency induces cell-specific changes in the steady-state transcript levels of genes
involved in folate metabolism and 1-carbon transfer reactions in human colonic epithelial
cells. J. Nutr. 137(3), 607613.
Holm, J., Hansen, S. I., Hoier-Madsen, M., and Bostad, L. (1991). High-affinity folate
binding in human choroid plexus. Characterization of radioligand binding, immunore-
activity, molecular heterogeneity and hydrophobic domain of the binding protein.
Biochem. J. 280(Pt 1), 267271.
Hooijberg, J. H., Broxterman, H. J., Kool, M., Assaraf, Y. G., Peters, G. J., Noordhuis, P.,
Scheper, R. J., Borst, P., Pinedo, H. M., and Jansen, G. (1999). Antifolate resistance
mediated by the multidrug resistance proteins MRP1 and MRP2. Cancer Res. 59(11),
25322535.
Hooijberg, J. H., Jansen, G., Assaraf, Y. G., Kathmann, I., Pieters, R., Laan, A. C.,
Veerman, A. J., Kaspers, G. J., and Peters, G. J. (2004). Folate concentration dependent
transport activity of the Multidrug Resistance Protein 1 (ABCC1). Biochem. Pharmacol. 67
(8), 15411548.
Hooijberg, J. H., Peters, G. J., Assaraf, Y. G., Kathmann, I., Priest, D. G., Bunni, M. A.,
Veerman, A. J., Scheffer, G. L., Kaspers, G. J., and Jansen, G. (2003). The role of
multidrug resistance proteins MRP1, MRP2 and MRP3 in cellular folate homeostasis.
Biochem. Pharmacol. 65(5), 765771.
Hopper, E., Belinsky, M. G., Zeng, H., Tosolini, A., Testa, J. R., and Kruh, G. D. (2001).
Analysis of the structure and expression pattern of MRP7 (ABCC10), a new member of
the MRP subfamily. Cancer Lett. 162(2), 181191.
Horns, R. C., Jr, Dower, W. J., and Schimke, R. T. (1984). Gene amplification in a
leukemic patient treated with methotrexate. J. Clin. Oncol. 2(1), 27.
Ifergan, I., Jansen, G., and Assaraf, Y. G. (2005). Cytoplasmic confinement of breast cancer
resistance protein (BCRP/ABCG2) as a novel mechanism of adaptation to short-term
folate deprivation. Mol. Pharmacol. 67(4), 13491359.
Adaptation to Folate Deprivation 135
Ifergan, I., Shafran, A., Jansen, G., Hooijberg, J. H., Scheffer, G. L., and Assaraf, Y. G.
(2004). Folate deprivation results in the loss of breast cancer resistance protein (BCRP/
ABCG2) expression. A role for BCRP in cellular folate homeostasis. J. Biol. Chem.
279(24), 2552725534.
Ishikawa, T. (1992). The ATP-dependent glutathione S-conjugate export pump. Trends
Biochem. Sci. 17(11), 463468.
Jackson, R. C., and Harrap, K. R. (1973). Studies with a mathematical model of folate
metabolism. Arch. Biochem. Biophys. 158(2), 827841.
James, S. J., Basnakian, A. G., and Miller, B. J. (1994). In vitro folate deficiency induces
deoxynucleotide pool imbalance, apoptosis, and mutagenesis in Chinese hamster ovary
cells. Cancer Res. 54(19), 50755080.
Jansen, G., Westerhof, G. R., Jarmuszewski, M. J., Kathmann, I., Rijksen, G., and
Schornagel, J. H. (1990). Methotrexate transport in variant human CCRF-CEM leuke-
mia cells with elevated levels of the reduced folate carrier. Selective effect on carrier-
mediated transport of physiological concentrations of reduced folates. J. Biol. Chem.
265(30), 1827218277.
Joelson, D. W., Fiebig, E. W., and Wu, A. H. (2007). Diminished need for folate
measurements among indigent populations in the post folic acid supplementation era.
Arch. Pathol. Lab. Med. 131(3), 477480.
Jonker, J. W., Buitelaar, M., Wagenaar, E., Van Der Valk, M. A., Scheffer, G. L.,
Scheper, R. J., Plosch, T., Kuipers, F., Elferink, R. P., Rosing, H., Beijnen, J. H., and
Schinkel, A. H. (2002). The breast cancer resistance protein protects against a major
chlorophyll-derived dietary phototoxin and protoporphyria. Proc. Natl. Acad. Sci. USA
99(24), 1564915654.
Jonker, J. W., Merino, G., Musters, S., van Herwaarden, A. E., Bolscher, E., Wagenaar, E.,
Mesman, E., Dale, T. C., and Schinkel, A. H. (2005). The breast cancer resistance
protein BCRP (ABCG2) concentrates drugs and carcinogenic xenotoxins into milk.
Nat. Med. 11(2), 127129.
Jukes, T. H. (1987). Searching for magic bullets: early approaches to chemotherapy-anti-
folates, methotrexatethe Bruce F. Cain memorial award lecture. Cancer Res. 47(21),
55285536.
Kane, M. A., Elwood, P. C., Portillo, R. M., Antony, A. C., Najfeld, V., Finley, A.,
Waxman, S., and Kolhouse, J. F. (1988). Influence on immunoreactive folate-binding
proteins of extracellular folate concentration in cultured human cells. J. Clin. Invest.
81(5), 13981406.
Keppler, D., Leier, I., and Jedlitschky, G. (1997). Transport of glutathione conjugates and
glucuronides by the multidrug resistance proteins MRP1 and MRP2. Biol. Chem. 378(8),
787791.
Kim, Y. I. (2004). Folate and DNA methylation: a mechanistic link between folate
deficiency and colorectal cancer? Cancer Epidemiol. Biomarkers Prev. 13(4), 511519.
Kim, Y. I. (2005). Nutritional epigenetics: Impact of folate deficiency on DNA methylation
and colon cancer susceptibility. J. Nutr. 135(11), 27032709.
Konig, P., Giraldo, R., Chapman, L., and Rhodes, D. (1996). The crystal structure of the
DNA-binding domain of yeast RAP1 in complex with telomeric DNA. Cell 85(1),
125136.
Krishnamurthy, P., and Schuetz, J. D. (2006). Role of ABCG2/BCRP in biology and
medicine. Annu. Rev. Pharmacol. Toxicol. 46, 381410.
Kruh, G. D., and Belinsky, M. G. (2003). The MRP family of drug efflux pumps. Oncogene
22(47), 75377552.
Kusuhara, H., Han, Y. H., Shimoda, M., Kokue, E., Suzuki, H., and Sugiyama, Y. (1998).
Reduced folate derivatives are endogenous substrates for cMOAT in rats. Am J. Physiol
275(4 Pt 1), G789G796.
136 Ilan Ifergan and Yehuda G. Assaraf
Lacey, S. W., Sanders, J. M., Rothberg, K. G., Anderson, R. G., and Kamen, B. A. (1989).
Complementary DNA for the folate binding protein correctly predicts anchoring to the
membrane by glycosyl-phosphatidylinositol. J. Clin. Invest. 84(2), 715720.
Lamers, Y., Prinz-Langenohl, R., Bramswig, S., and Pietrzik, K. (2006). Red blood cell folate
concentrations increase more after supplementation with [6S]-5-methyltetrahydrofolate
than with folic acid in women of childbearing age. Am. J. Clin. Nutr. 84(1), 156161.
Laurence, K. M., James, N., Miller, M. H., Tennant, G. B., and Campbell, H. (1981).
Double-blind randomised controlled trial of folate treatment before conception to prevent
recurrence of neural-tube defects. Br. Med. J. (Clin. Res. Ed.) 282(6275), 15091511.
Lee, H. C., Shoda, R., Krall, J. A., Foster, J. D., Selhub, J., and Rosenberry, T. L. (1992).
Folate binding protein from kidney brush border membranes contains components
characteristic of a glycoinositol phospholipid anchor. Biochemistry 31(12), 32363243.
Liani, E., Rothem, L., Bunni, M. A., Smith, C. A., Jansen, G., and Assaraf, Y. G. (2003).
Loss of folylpoly-gamma-glutamate synthetase activity is a dominant mechanism of
resistance to polyglutamylation-dependent novel antifolates in multiple human leukemia
sublines. Int. J. Cancer 103(5), 587599.
Lin, B. F., Huang, R. F., and Shane, B. (1993). Regulation of folate and one-carbon
metabolism in mammalian cells. III. Role of mitochondrial folylpoly-gamma-glutamate
synthetase. J. Biol. Chem. 268(29), 2167421679.
Liu, X. Y., and Matherly, L. H. (2002). Analysis of membrane topology of the human
reduced folate carrier protein by hemagglutinin epitope insertion and scanning glycosyl-
ation insertion mutagenesis. Biochim. Biophys. Acta 1564(2), 333342.
Loo, T. W., and Clarke, D. M. (1994). Mutations to amino acids located in predicted
transmembrane segment 6 (TM6) modulate the activity and substrate specificity of
human P-glycoprotein. Biochemistry 33(47), 1404914057.
Lubhy, A. L., Eagle, F. J., Roth, E., and Cooperman, J. M. (1961). Relapsing megaloblastic
anemia in an infant due to a specific defect in gastrointestinal absorption of folic acid.
Am. J. Dis. Child 102, 482483.
Ma, L., Krishnamachary, N., and Center, M. S. (1995). Phosphorylation of the multidrug
resistance associated protein gene encoded protein P190. Biochemistry 34(10), 33383343.
Maliepaard, M., van Gastelen, M. A., Tohgo, A., Hausheer, F. H., van Waardenburg, R. C.,
de Jong, L. A., Pluim, D., Beijnen, J. H., and Schellens, J. H. (2001). Circumvention of
breast cancer resistance protein (BCRP)-mediated resistance to camptothecins in vitro using
non-substrate drugs or the BCRP inhibitor GF120918. Clin. Cancer Res. 7(4), 935941.
Mason, J. B., and Rosenberg, I. H. (1994). Intestinal absorption of folate. In Physiology
of the Gastrointestinal Tract (L. R. Johnson, ed.), pp. 19791995. Raven Press,
New York.
Matherly, L. H., and Goldman, D. I. (2003). Membrane transport of folates. Vitam. Horm.
66, 403456.
McEwan, G. T., Lucas, M. L., Denvir, M., Raj, M., McColl, K. E., Russell, R. I., and
Mathan, V. I. (1990). A combined TDDA-PVC pH and reference electrode for use in
the upper small intestine. J. Med. Eng. Technol. 14(1), 1620.
McGuire, J. J., Russell, C. A., and Balinska, M. (2000). Human cytosolic and mitochondrial
folylpolyglutamate synthetase are electrophoretically distinct. Expression in antifolate-
sensitive and -resistant human cell lines. J. Biol. Chem. 275(17), 1301213016.
McHugh, M., and Cheng, Y. C. (1979). Demonstration of a high affinity folate binder in
human cell membranes and its characterization in cultured human KB cells. J. Biol. Chem.
254(22), 1131211318.
Melnyk, S., Pogribna, M., Miller, B. J., Basnakian, A. G., Pogribny, I. P., and James, S. J.
(1999). Uracil misincorporation, DNA strand breaks, and gene amplification are asso-
ciated with tumorigenic cell transformation in folate deficient/repleted Chinese hamster
ovary cells. Cancer Lett. 146(1), 3544.
Adaptation to Folate Deprivation 137
Miyake, K., Mickley, L., Litman, T., Zhan, Z., Robey, R., Cristensen, B., Brangi, M.,
Greenberger, L., Dean, M., Fojo, T., and Bates, S. E. (1999). Molecular cloning of
cDNAs which are highly overexpressed in mitoxantrone-resistant cells: Demonstration
of homology to ABC transport genes. Cancer Res. 59(1), 813.
Mogi, M., Yang, J., Lambert, J. F., Colvin, G. A., Shiojima, I., Skurk, C., Summer, R.,
Fine, A., Quesenberry, P. J., and Walsh, K. (2003). Akt signaling regulates side popula-
tion cell phenotype via Bcrp1 translocation. J. Biol. Chem. 278(40), 3906839075.
Morgillo, F., and Lee, H. Y. (2005). Resistance to epidermal growth factor receptor-targeted
therapy. Drug Resist. Updat. 8(5), 298310.
Nguyen, T. T., Dyer, D. L., Dunning, D. D., Rubin, S. A., Grant, K. E., and Said, H. M.
(1997). Human intestinal folate transport: Cloning, expression, and distribution of
complementary RNA. Gastroenterology 112(3), 783791.
Nozawa, T., Imai, K., Nezu, J., Tsuji, A., and Tamai, I. (2004). Functional characterization
of pH-sensitive organic anion transporting polypeptide OATP-B in human. J. Pharmacol
Exp. Ther. 308(2), 438445.
Ozaki, Y., King, R. W., and Carey, P. R. (1981). Methotrexate and folate binding to
dihydrofolate reductase. Separate characterization of the pteridine and p-aminobenzoyl
binding sites by resonance Raman spectroscopy. Biochemistry 20(11), 32193225.
Ozvegy, C., Litman, T., Szakacs, G., Nagy, Z., Bates, S., Varadi, A., and Sarkadi, B. (2001).
Functional characterization of the human multidrug transporter, ABCG2, expressed in
insect cells. Biochem. Biophys. Res. Commun. 285(1), 111117.
Pathak, P., Kapil, U., Kapoor, S. K., Saxena, R., Kumar, A., Gupta, N., Dwivedi, S. N.,
Singh, R., and Singh, P. (2004). Prevalence of multiple micronutrient deficiencies
amongst pregnant women in a rural area of Haryana. Indian J. Pediatr. 71(11), 10071014.
Paulusma, C. C., Bosma, P. J., Zaman, G. J., Bakker, C. T., Otter, M., Scheffer, G. L.,
Scheper, R. J., Borst, P., and Oude Elferink, R. P. (1996). Congenital jaundice in rats
with a mutation in a multidrug resistance-associated protein gene. Science 271(5252),
11261128.
Piedrahita, J. A., Oetama, B., Bennett, G. D., van Waes, J., Kamen, B. A., Richardson, J.,
Lacey, S. W., Anderson, R. G., and Finnell, R. H. (1999). Mice lacking the folic acid-
binding protein Folbp1 are defective in early embryonic development. Nat. Genet. 23(2),
228232.
Polgar, O., Ozvegy-Laczka, C., Robey, R. W., Morisaki, K., Okada, M., Tamaki, A.,
Koblos, G., Elkind, N. B., Ward, Y., Dean, M., Sarkadi, B., and Bates, S. E. (2006).
Mutational studies of G553 in TM5 of ABCG2: A residue potentially involved in
dimerization. Biochemistry 45(16), 52515260.
Polgar, O., Robey, R. W., Morisaki, K., Dean, M., Michejda, C., Sauna, Z. E.,
Ambudkar, S. V., Tarasova, N., and Bates, S. E. (2004). Mutational analysis of
ABCG2: role of the GXXXG motif. Biochemistry 43(29), 94489456.
Poncz, M., and Cohen, A. (1996). Long-term treatment of congenital folate malabsorption.
J. Pediatr. 129(6), 948.
Qiu, A., Jansen, M., Sakaris, A., Min, S. H., Chattopadhyay, S., Tsai, E., Sandoval, C.,
Zhao, R., Akabas, M. H., and Goldman, I. D. (2006). Identification of an intestinal folate
transporter andthe molecular basis for hereditaryfolate malabsorption. Cell 127(5), 917928.
Qureshi, A. A., Rosenblatt, D. S., and Cooper, B. A. (1994). Inherited disorders of
cobalamin metabolism. Crit. Rev. Oncol. Hematol. 17(2), 133151.
Ramaekers, V. T., and Blau, N. (2004). Cerebral folate deficiency. Dev. Med. Child. Neurol.
46(12), 843851.
Ramaekers, V. T., Hausler, M., Opladen, T., Heimann, G., and Blau, N. (2002). Psycho-
motor retardation, spastic paraplegia, cerebellar ataxia and dyskinesia associated with low
5-methyltetrahydrofolate in cerebrospinal fluid: A novel neurometabolic condition
responding to folinic acid substitution. Neuropediatrics 33(6), 301308.
138 Ilan Ifergan and Yehuda G. Assaraf
Ramaekers, V. T., Rothenberg, S. P., Sequeira, J. M., Opladen, T., Blau, N.,
Quadros, E. V., and Selhub, J. (2005). Autoantibodies to folate receptors in the cerebral
folate deficiency syndrome. N. Engl. J. Med. 352(19), 19851991.
Ramakrishnan, S., Sulochana, K. N., Lakshmi, S., Selvi, R., and Angayarkanni, N. (2006).
Biochemistry of homocysteine in health and diseases. Indian J. Biochem. Biophys. 43(5),
275283.
Rasouli, M. L., Nasir, K., Blumenthal, R. S., Park, R., Aziz, D. C., and Budoff, M. J.
(2005). Plasma homocysteine predicts progression of atherosclerosis. Atherosclerosis
181(1), 159165.
Ratnam, M., and Freisheim, J. H. (1992). Proteins involved in the transport of folates and
antifolates by normal and neoplastic cells. In Folic Acid Metabolism in Health and
Disease (M. F. Picciano, E. L. R. Stokstad, and J. F. Gregory, eds.), pp. 91120. Wiley-
Liss, Inc, New York.
Ratnam, M., Marquardt, H., Duhring, J. L., and Freisheim, J. H. (1989). Homologous
membrane folate binding proteins in human placenta: Cloning and sequence of a cDNA.
Biochemistry 28(20), 82498254.
Richards, R. G., Brown, O. E., Gillison, M. L., and Sedwick, W. D. (1986). Drug
concentration-dependent DNA lesions are induced by the lipid-soluble antifolate, piri-
trexim (BW301U). Mol. Pharmacol. 30(6), 651658.
Rothem, L., Ifergan, I., Kaufman, Y., Priest, D. G., Jansen, G., and Assaraf, Y. G. (2002).
Resistance to multiple novel antifolates is mediated via defective drug transport resulting
from clustered mutations in the reduced folate carrier gene in human leukaemia cell lines.
Biochem. J. 367(Pt 3), 741750.
Rothenberg, S. P., da Costa, M. P., Sequeira, J. M., Cracco, J., Roberts, J. L., Weedon, J.,
and Quadros, E. V. (2004). Autoantibodies against folate receptors in women with a
pregnancy complicated by a neural-tube defect. N. Engl. J. Med. 350(2), 134142.
Rubin, R. C., Henderson, E. S., Owens, E. S., and Rall, D. P. (1967). Uptake of
methotrexate-3H by rabbit kidney slices. Cancer Res. 27(3), 553557.
Sadasivan, E., and Rothenberg, S. P. (1989). The complete amino acid sequence of a human
folate binding protein from KB cells determined from the cDNA. J. Biol. Chem. 264(10),
58065811.
Saier, M. H., Jr, Beatty, J. T., Goffeau, A., Harley, K. T., Heijne, W. H., Huang, S. C.,
Jack, D. L., Jahn, P. S., Lew, K., Liu, J., Pao, S. S., Paulsen, I. T., et al. (1999). The major
facilitator superfamily. J. Mol. Microbiol. Biotechnol. 1(2), 257279.
Sarkadi, B., Muller, M., Homolya, L., Hollo, Z., Seprodi, J., Germann, U. A.,
Gottesman, M. M., Price, E. M., and Boucher, R. C. (1994). Interaction of bioactive
hydrophobic peptides with the human multidrug transporter. Faseb. J. 8(10), 766770.
Sauls, D. L., Arnold, E. K., Bell, C. W., Allen, J. C., and Hoffman, M. (2007). Pro-
thrombotic and pro-oxidant effects of diet-induced hyperhomocysteinemia. Thromb.
Res. 120(1), 117126.
Sauls, D. L., Boyd, L. C., Allen, J. C., and Hoffman, M. (2004). Differences in the metabolic
response to exogenous homocysteine in juvenile and adult rabbits. J. Nutr. Biochem. 15
(2), 96102.
Sauls, D. L., Lockhart, E., Warren, M. E., Lenkowski, A., Wilhelm, S. E., and Hoffman, M.
(2006). Modification of fibrinogen by homocysteine thiolactone increases resistance to
fibrinolysis: A potential mechanism of the thrombotic tendency in hyperhomocysteine-
mia. Biochemistry 45(8), 24802487.
Schimke, R. T. (1984). Gene amplification in cultured animal cells. Cell 37(3), 705713.
Schinkel, A. H., Mayer, U., Wagenaar, E., Mol, C. A., van Deemter, L., Smit, J. J., van der
Valk, M. A., Voordouw, A. C., Spits, H., van Tellingen, O., Zijlmans, J. M., Fibbe, W. E.,
et al. (1997). Normal viability and altered pharmacokinetics in mice lacking mdr1-type
(drug-transporting) P-glycoproteins. Proc. Natl. Acad. Sci. USA 94(8), 40284033.
Adaptation to Folate Deprivation 139
Schmitt, L., and Tampe, R. (2000). Affinity, specificity, diversity: A challenge for the ABC
transporter TAP in cellular immunity. Chembiochem. 1(1), 1635.
Scott, J. M. (1999). Folate and vitamin B12. Proc. Nutr. Soc. 58(2), 441448.
Seither, R. L., Trent, D. F., Mikulecky, D. C., Rape, T. J., and Goldman, I. D. (1989).
Folate-pool interconversions and inhibition of biosynthetic processes after exposure of
L1210 leukemia cells to antifolates. Experimental and network thermodynamic analyses
of the role of dihydrofolate polyglutamylates in antifolate action in cells. J. Biol. Chem.
264(29), 1701617023.
Selhub, J., Dhar, G. J., and Rosenberg, I. H. (1986). Gastrointestinal absorption of antifolates.
Int. Encyl. Pharmacol. Ther. Vol. 118, pp. 147156. Pergamon, Oxford/NewYork.
Selhub, J., and Rosenberg, I. H. (1978). Demonstration of high-affinity folate binding
activity associated with the brush border membranes of rat kidney. Proc. Natl. Acad. Sci.
USA 75(7), 30903093.
Selhub, J., and Rosenberg, I. H. (1981). Folate transport in isolated brush border membrane
vesicles from rat intestine. J. Biol. Chem. 256(9), 44894493.
Seller, M. J., and Nevin, N. C. (1984). Periconceptional vitamin supplementation and the
prevention of neural tube defects in south-east England and Northern Ireland. J. Med.
Genet. 21(5), 325330.
Shane, B. (1989). Folylpolyglutamate synthesis and role in the regulation of one-carbon
metabolism. Vitam. Horm. 45, 263335.
Shayeghi, M., Latunde-Dada, G. O., Oakhill, J. S., Laftah, A. H., Takeuchi, K.,
Halliday, N., Khan, Y., Warley, A., McCann, F. E., Hider, R. C., Frazer, D. M.,
Anderson, G. J., et al. (2005). Identification of an intestinal heme transporter. Cell 122
(5), 789801.
Shen, F., Ross, J. F., Wang, X., and Ratnam, M. (1994). Identification of a novel folate
receptor, a truncated receptor, and receptor type beta in hematopoietic cells: cDNA
cloning, expression, immunoreactivity, and tissue specificity. Biochemistry 33(5),
12091215.
Shen, F., Wu, M., Ross, J. F., Miller, D., and Ratnam, M. (1995). Folate receptor
type gamma is primarily a secretory protein due to lack of an efficient signal for
glycosylphosphatidylinositol modification: Protein characterization and cell type
specificity. Biochemistry 34(16), 56605665.
Sierra, E. E., Brigle, K. E., Spinella, M. J., and Goldman, I. D. (1997). pH dependence of
methotrexate transport by the reduced folate carrier and the folate receptor in L1210
leukemia cells. Further evidence for a third route mediated at low pH. Biochem. Pharma-
col. 53(2), 223231.
Sirotnak, F. M. (1985). Obligate genetic expression in tumor cells of a fetal membrane
property mediating folate. transport: Biological significance and implications for
improved therapy of human cancer. Cancer Res. 45(9), 39924000.
Sirotnak, F. M. (1987). Determinants of resistance to antifolates: Biochemical phenotypes,
their frequency of occurrence and circumvention. NCI Monogr. Volume 1987, (Number
5), 2735.
Sirotnak, F. M., and Tolner, B. (1999). Carrier-mediated membrane transport of folates in
mammalian cells. Annu. Rev. Nutr. 19, 91122.
Smit, J. W., Huisman, M. T., van Tellingen, O., Wiltshire, H. R., and Schinkel, A. H.
(1999). Absence or pharmacological blocking of placental P-glycoprotein profoundly
increases fetal drug exposure. J. Clin. Invest. 104(10), 14411447.
Smithells, R. W., Nevin, N. C., Seller, M. J., Sheppard, S., Harris, R., Read, A. P.,
Fielding, D. W., Walker, S., Schorah, C. J., and Wild, J. (1983). Further experience of
vitamin supplementation for prevention of neural tube defect recurrences. Lancet 1
(8332), 10271031.
140 Ilan Ifergan and Yehuda G. Assaraf
Smithells, R. W., Sheppard, S., and Schorah, C. J. (1976). Vitamin dificiencies and neural
tube defects. Arch. Dis. Child 51(12), 944950.
Smithells, R. W., Sheppard, S., Schorah, C. J., Seller, M. J., Nevin, N. C., Harris, R.,
Read, A. P., and Fielding, D. W. (1981). Apparent prevention of neural tube defects by
periconceptional vitamin supplementation. Arch. Dis. Child 56(12), 911918.
Sparreboom, A., van Asperen, J., Mayer, U., Schinkel, A. H., Smit, J. W., Meijer, D. K.,
Borst, P., Nooijen, W. J., Beijnen, J. H., and van Tellingen, O. (1997). Limited oral
bioavailability and active epithelial excretion of paclitaxel (Taxol) caused by P-glycopro-
tein in the intestine. Proc. Natl. Acad. Sci. USA 94(5), 20312035.
Spector, R., and Johanson, C. (2006). Micronutrient and urate transport in choroid plexus
and kidney: implications for drug therapy. Pharm. Res. 23(11), 25152524.
Spiegelstein, O., Eudy, J. D., and Finnell, R. H. (2000). Identification of two putative novel
folate receptor genes in humans and mouse. Gene 258(1-2), 117125.
Stanger, O. (2002). Physiology of folic acid in health and disease. Curr. Drug Metab. 3(2),
211223.
Stark, M., Rothem, L., Jansen, G., Scheffer, G. L., Goldman, I. D., and Assaraf, Y. G.
(2003). Antifolate resistance associated with loss of MRP1 expression and function in
Chinese hamster ovary cells with markedly impaired export of folate and cholate.
Mol. Pharmacol. 64(2), 220227.
Starkebaum, G., and Harlan, J. M. (1986). Endothelial cell injury due to copper-catalyzed
hydrogen peroxide generation from homocysteine. J. Clin. Invest. 77(4), 13706.
Stokstad, E. L. R. (1990). In Folic acid metabolism in health and disease (M. F. Picciano,
E. L. R. Stokstad, and J. F. Greogory, eds.), pp. 121. Wiley-Liss, New York.
Subramanian, V. S., Chatterjee, N., and Said, H. M. (2003). Folate uptake in the human
intestine: Promoter activity and effect of folate deficiency. J. Cell. Physiol. 196(2),
403408.
Sun, X. L., and Antony, A. C. (1996). Evidence that a specific interaction between an 18-
base cis-element in the 5-untranslated region of human folate receptor-alpha mRNA
and a 46-kDa cytosolic trans-factor is critical for translation. J. Biol. Chem. 271(41),
2553925547.
Szakacs, G., Paterson, J. K., Ludwig, J. A., Booth-Genthe, C., and Gottesman, M. M.
(2006). Targeting multidrug resistance in cancer. Nat. Rev. Drug. Discov. 5(3), 219234.
Tammur, J., Prades, C., Arnould, I., Rzhetsky, A., Hutchinson, A., Adachi, M.,
Schuetz, J. D., Swoboda, K. J., Ptacek, L. J., Rosier, M., Dean, M., and Allikmets, R.
(2001). Two new genes from the human ATP-binding cassette transporter superfamily,
ABCC11 and ABCC12, tandemly duplicated on chromosome 16q12. Gene 273(1),
8996.
Tannock, I. F., and Rotin, D. (1989). Acid pH in tumors and its potential for therapeutic
exploitation. Cancer Res. 49(16), 43734384.
Trent, J. M., Buick, R. N., Olson, S., Horns, R. C., Jr, and Schimke, R. T. (1984).
Cytologic evidence for gene amplification in methotrexate-resistant cells obtained from
a patient with ovarian adenocarcinoma. J. Clin. Oncol. 2(1), 815.
Ubbink, J. B., Vermaak, W. J., van der Merwe, A., and Becker, P. J. (1993). Vitamin B-12,
vitamin B-6, and folate nutritional status in men with hyperhomocysteinemia. Am. J.
Clin. Nutr. 57(1), 4753.
van Herwaarden, A. E., Jonker, J. W., Wagenaar, E., Brinkhuis, R. F., Schellens, J. H.,
Beijnen, J. H., and Schinkel, A. H. (2003). The breast cancer resistance protein (Bcrp1/
Abcg2) restricts exposure to the dietary carcinogen 2-amino-1-methyl-6-phenylimidazo
[4,5-b]pyridine. Cancer Res. 63(19), 64476452.
van Herwaarden, A. E., Wagenaar, E., Merino, G., Jonker, J. W., Rosing, H.,
Beijnen, J. H., and Schinkel, A. H. (2007). Multidrug transporter ABCG2/breast cancer
Adaptation to Folate Deprivation 141
resistance protein secretes riboflavin (vitamin B2) into milk. Mol. Cell Biol. 27(4),
12471253.
Vergel, R. G., Sanchez, L. R., Heredero, B. L., Rodriguez, P. L., and Martinez, A. J. (1990).
Primary prevention of neural tube defects with folic acid supplementation: Cuban
experience. Prenat. Diagn. 10(3), 149152.
Volk, E. L., and Schneider, E. (2003). Wild-type breast cancer resistance protein (BCRP/
ABCG2) is a methotrexate polyglutamate transporter. Cancer Res. 63(17), 55385543.
Walling, J. (2006). From methotrexate to pemetrexed and beyond. A review of the
pharmacodynamic and clinical properties of antifolates. Invest. New Drugs 24(1), 3777.
Wang, X., Shen, F., Freisheim, J. H., Gentry, L. E., and Ratnam, M. (1992). Differential
stereospecificities and affinities of folate receptor isoforms for folate compounds and
antifolates. Biochem. Pharmacol. 44(9), 18981901.
Wang, Y., Zhao, R., and Goldman, I. D. (2004). Characterization of a folate transporter in
HeLa cells with a low pH optimum and high affinity for pemetrexed distinct from the
reduced folate carrier. Clin. Cancer Res. 10(18 Pt 1), 62566264.
Weitman, S. D., Weinberg, A. G., Coney, L. R., Zurawski, V. R., Jennings, D. S., and
Kamen, B. A. (1992). Cellular localization of the folate receptor: potential role in drug
toxicity and folate homeostasis. Cancer Res. 52(23), 67086711.
Wielinga, P., Hooijberg, J. H., Gunnarsdottir, S., Kathmann, I., Reid, G., Zelcer, N., van
der Born, K., de Haas, M., van der Heijden, I., Kaspers, G., Wijnholds, J., Jansen, G.,
et al. (2005). The human multidrug resistance protein MRP5 transports folates and can
mediate cellular resistance against antifolates. Cancer Res. 65(10), 44254430.
Wijnholds, J., deLange, E. C., Scheffer, G. L., van den Berg, D. J., Mol, C. A., van der
Valk, M., Schinkel, A. H., Scheper, R. J., Breimer, D. D., and Borst, P. (2000).
Multidrug resistance protein 1 protects the choroid plexus epithelium and contributes
to the blood-cerebrospinal fluid barrier. J. Clin. Invest. 105(3), 279285.
Williams, F. M., and Flintoff, W. F. (1998). Structural organization of the human reduced
folate carrier gene: Evidence for 5
0
heterogeneity in lymphoblast mRNA. Somat. Cell
Mol. Genet. 24(3), 143156.
Xiao, X., Tang, Y. S., Mackins, J. Y., Sun, X. L., Jayaram, H. N., Hansen, D. K., and
Antony, A. C. (2001). Isolation and characterization of a folate receptor mRNA-binding
trans-factor from human placenta. Evidence favoring identity with heterogeneous
nuclear ribonucleoprotein E1. J. Biol. Chem. 276(44), 4151041517.
Young, L., Leonhard, K., Tatsuta, T., Trowsdale, J., and Langer, T. (2001). Role of the
ABC transporter Mdl1 in peptide export from mitochondria. Science 291(5511),
21352138.
Zeng, H., Bain, L. J., Belinsky, M. G., and Kruh, G. D. (1999). Expression of multidrug
resistance protein-3 (multispecific organic anion transporter-D) in human embryonic
kidney 293 cells confers resistance to anticancer agents. Cancer Res. 59(23), 59645967.
Zeng, H., Chen, Z. S., Belinsky, M. G., Rea, P. A., and Kruh, G. D. (2001). Transport of
methotrexate (MTX) and folates by multidrug resistance protein (MRP) 3 and MRP1:
Effect of polyglutamylation on MTX transport. Cancer Res. 61(19), 72257232.
Zhang, L., Wong, S. C., and Matherly, L. H. (1998). Transcript heterogeneity of the human
reduced folate carrier results from the use of multiple promoters and variable splicing of
alternative upstream exons. Biochem. J. 332(Pt 3), 773780.
Zhao, R., Gao, F., and Goldman, I. D. (2002). Reduced folate carrier transports thiamine
monophosphate: An alternative route for thiamine delivery into mammalian cells. Am. J.
Physiol. Cell Physiol. 282(6), C1512C1517.
Zhao, R., Min, S. H., Qiu, A., Sakaris, A., Goldberg, G. L., Sandoval, C., Malatack, J. J.,
Rosenblatt, D. S., and Goldman, I. D. (2007). The spectrum of mutations in the PCFT
gene, coding for an intestinal folate transporter, that are the basis for hereditary folate
malabsorption. Blood 110, 11471152.
142 Ilan Ifergan and Yehuda G. Assaraf
Zhao, R., Russell, R. G., Wang, Y., Liu, L., Gao, F., Kneitz, B., Edelmann, W., and
Goldman, I. D. (2001). Rescue of embryonic lethality in reduced folate carrier-deficient
mice by maternal folic acid supplementation reveals early neonatal failure of hemato-
poietic organs. J. Biol. Chem. 276(13), 102241028.
Zhu, H., Wlodarczyk, B. J., Scott, M., Yu, W., Merriweather, M., Gelineau-van Waes, J.,
Schwartz, R. J., and Finnell, R. H. (2007). Cardiovascular abnormalities in Folr1
knockout mice and folate rescue. Birth Defects Res. A Clin. Mol. Teratol. 79(4), 25768.
Zhu, W. Y., Alliegro, M. A., and Melera, P. W. (2001). The rate of folate receptor alpha
(FR alpha) synthesis in folate depleted CHL cells is regulated by a translational mecha-
nism sensitive to media folate levels, while stable overexpression of its mRNA is mediated
by gene amplification and an increase in transcript half-life. J. Cell. Biochem. 81(2),
205219.
Zhu, W. Y., Bunni, M., Priest, D. G., DiCapua, J. L., Dressler, J. M., Chen, Z., and
Melera, P. W. (2002). Severe folate restriction results in depletion of and alteration in the
composition of the intracellular folate pool, moderate sensitization to methotrexate and
trimetrexate, upregulation of endogenous DHFR activity, and overexpression of metal-
lothionein II and folate receptor alpha that, upon folate repletion, confer drug resistance
to CHL cells. J. Exp. Ther. Oncol. 2(5), 264277.
Adaptation to Folate Deprivation 143
C H A P T E R F I V E
Structure and Function of the
Reduced Folate Carrier: A Paradigm
of a Major Facilitator Superfamily
Mammalian Nutrient Transporter
Larry H. Matherly*
,
and Zhanjun Hou
Contents
I. Introduction 146
II. MFS of Transporters 147
III. Folate Transport in Tissue Folate Homeostasis and Physiology:
Role of Multiple Transport Systems for Folate Uptake and Efflux 150
IV. Role of RFC in Antifolate Chemotherapy 152
V. Functional Properties of RFC 154
VI. Biochemistry of RFC 156
VII. Cloning of RFC cDNAs That Restore Transport to
Transport-Impaired Cultured Cells 158
VIII. Topological Structure of RFC 159
A. Topological structure by HA epitope accessibility
and N-glycosylation scanning mutagenesis 159
B. Topological structure by scanning cysteine
accessibility methods 163
IX. Insights into Structural and Functional Determinants of RFC from
Studies of Mutant RFC Proteins 164
A. Identification of structurally or functionally important amino
acids by selecting for mutant RFCs with antifolate inhibitors 165
B. Identification of structurally or functionally important
amino acids in RFC by homology comparisons and
site-directed mutagenesis 166
C. Deletional and insertional mutagenesis of RFC 167
D. Localization of a substrate-binding domain
by radioaffinity labeling 168
Vitamins and Hormones, Volume 79
#
2008 Elsevier Inc.
ISSN 0083-6729, DOI: 10.1016/S0083-6729(08)00405-6 All rights reserved.
* Developmental Therapeutics Program, Barbara Ann Karmanos Cancer Institute, Wayne State University
School of Medicine, Detroit, Michigan 48201
{
Developmental Therapeutics Program, Barbara Ann Karmanos Cancer Institute, Wayne State University
School of Medicine, Detroit, Michigan 48201
145
E. SCAM for mapping the substrate translocation pathway 169
F. Mapping helix packing associations in hRFC 172
X. Conclusions 173
Acknowledgment 175
References 175
Abstract
Folates are essential for life and folate deficiency contributes to a host of health
problems including cardiovascular disease, fetal abnormalities, neurological
disorders, and cancer. Antifolates, represented by methotrexate, continue to
occupy a unique niche among the modern day pharmacopoeia for cancer along
with other pathological conditions. This article focuses on the biology of the
membrane transport system termed the reduced folate carrier or RFC with a
particular emphasis on RFC structure and function. The ubiquitously expressed
RFCis the major transporter for folates inmammaliancells andtissues. Loss of RFC
expression or function portends potentially profound physiological or develop-
mental consequences. For chemotherapeutic antifolates used for cancer, loss of
RFCexpressionor synthesisof mutant RFCproteinwithimpairedfunctionresults in
antifolate resistance due to incomplete inhibition of cellular enzyme targets and
low levels of substrate for polyglutamate synthesis. The functional properties for
RFC were first documented nearly 40 years ago in murine leukemia cells. Since
1994, when RFCwas first cloned, tremendous advances inthe molecular biology of
RFCandbiochemical approachesfor studying thestructure of polytopic membrane
proteins have led to an increasingly detailed picture of the molecular structure of
the carrier, including its membrane topology, its N-glycosylation, identification
of functionally and structurally important domains and amino acids, and helix
packing associations. Although no crystal structure for RFC is yet available,
biochemical and molecular studies, combined with homology modeling, based
on homologous bacterial major facilitator superfamily transporters such as LacY,
now permit the development of experimentally testable hypotheses designed to
establish RFC structure and mechanism. 2008 Elsevier Inc.
I. Introduction
Folate is the generic term for water-soluble members of the B class of
vitamins that are required for normal tissue growth and development. Folic
acid is the synthetic form of the metabolically important folates found in cells
that differ in the level of oxidation of the pteridine ring, the nature of the one-
carbon substituent at the N
5
and N
10
positions, and the extent of g-glutamate
conjugation (Stokstad, 1990). The biological importance of reduced folates
derives fromtheir essential roles in one-carbon transfer leading to thymidylate,
purine nucleotides, serine, and methionine, and in biological methylation
reactions from S-adenosyl methionine (Stokstad, 1990).
146 Larry H. Matherly and Zhanjun Hou
Folates are hydrophilic molecules that are anions at physiological pH and
thus cross biological membranes only poorly by diffusion. Reflecting this,
mammalian cells have evolved sophisticated uptake systems for facilitating
cellular uptake of folate cofactors (Matherly and Goldman, 2003). The
reduced folate carrier (RFC) or SLC19A1 is expressed ubiquitously and is
recognized to be the major transport system for folates in mammalian cells
and tissues (Matherly and Goldman, 2003). In addition to its generalized
role in folate transport, RFC performs certain specialized tissue functions
including absorption across intestinal/colonic epithelia (Balamurugan and
Said, 2006; Chiao et al., 1997; Said, 2004), transport across the basolateral
membrane of renal proximal tubules (Kneuer et al., 2005), transplacental
transport of folates (Sweiry and Yudlievich, 1985), and folate transport
across the bloodbrain barrier (Spector and Johanson, 2006). Reflecting
these important physiological roles, low levels of RFC can be envisaged to
contribute to a number of pathophysiological states associated with folate
deficiency, including cardiovascular disease, fetal abnormalities, neurological
disorders, and possibly cancer (Matherly, 2004). Susceptibilities to these
conditions could be exacerbated with folate deficiency. Importantly, mem-
brane transport by RFC is also important for the antitumor activities of
antifolate therapeutics used for cancer chemotherapy such as methotrexate
(MTX) and pemetrexed (Matherly et al., 2007; Fig. 5.1).
This chapter focuses on the biology of the RFC with a particular
emphasis on RFC structure and function. The functional properties for
RFC were first documented nearly 40 years ago in murine leukemia cells
(Goldman et al., 1968). However, it is only since the cloning of the rodent
RFCs in 1994 (Dixon et al., 1994; Williams et al., 1994) and the human
RFC (hRFC) in 1995 (Moscow et al., 1995; Prasad et al., 1995; Williams
and Flintoff, 1995; Wong et al., 1995), and the application of molecular
biology approaches to engineer RFC for biochemical studies, that a detailed
picture of the molecular structure of this physiologically important carrier
has emerged, including its membrane topology, its N-glycosylation, and
identification of functionally and structurally important domains and amino
acids. In this chapter, we review the considerable progress in this area,
drawing significantly from the literature for RFC since 1994, along with
structural inferences from structurally homologous bacterial major facilitator
superfamily (MFS) members for which crystal structures were recently
reported (Abramson et al., 2003; Huang et al., 2003; Yin et al., 2006).
II. MFS of Transporters
The MFS of transporters is represented in animals, plants, fungi, lower
eukaryotes, bacteria, and eukaryotic organelles and transports a diverse
assortment of substrates in a uniport, symport, or antiport fashion, including
Structure and Function of the Reduced Folate Carrier 147
Folic Acid
HN
N
N
N H
2
N
O
CH
2
HN
O
HN CH CH
2
CH
2
CO
2
H
C
CO
2
H
Methotrexate
N
N
N
N H
2
N
NH
2
CH
2
N H
3
C
O
HN CH CH
2
CH
2
CO
2
H
C
CO
2
H
Tomudex (Ratitrexed)
HN
N H
3
C
O
CH
2
N H
3
C
O
S
HN CH CH
2
CH
2
CO
2
H
C
CO
2
H
5-Formyl Tetrahydrofolate
HN
N
N
NH H
2
N
O
CH
2
HN CHO
O
HN CH CH
2
CH
2
CO
2
H
C
CO
2
H
5-Methyl Tetrahydrofolate
HN
N
N
NH H
2
N
O
CH
2
HN CH
3
O
HN CH CH
2
CH
2
CO
2
H
C
CO
2
H
Pemetrexed
HN
N NH H
2
N
O
CH
2
H
2
C
O
HN CH CH
2
CH
2
CO
2
H
C
CO
2
H
Figure 5.1 Transport substrates for the reduced folate carrier (RFC). Structures are shown for folate and antifolate substrates for RFC.
amino acids, neurotransmitters, sugars, vitamins, nucleosides, and organic
phosphates (Saier et al., 1999). The MFS includes over 2000 sequenced
members (Chang et al., 2004), making it the largest secondary active
transporter family. MFS proteins typically contain 400600 amino acids
and a structural motif composed of two halves, each half composed of six
transmembrane-spanning a-helices connected by a large hydrophilic loop,
with cytosolic N- and C-termini.
Protein structural information is a prerequisite for understanding the
mechanism of membrane transport. It is remarkable that, given their abun-
dance and biological importance, there is a paucity of structural information
on the MFS proteins. This in part reflects difficulties in isolating sufficient
quantities of purified proteins and in crystallizing membrane proteins for
X-ray diffraction. Accordingly, structural insights have relied on an exten-
sive array of sophisticated biochemical and biophysical approaches, along
with homology structural modeling from the crystal structures that have
been reported.
To date, X-ray crystal structures have been reported for four MFS
proteins. The first structural evidence approaching atomic resolution was
for the oxalate transporter (OxlT) at 6.5 A
, generated by cryoelectron
microscopy (Hirai et al., 2002, 2003). In 2003, X-ray crystallographic
structures of the MFS proteins, the lactose/proton symporter (LacY)
(Abramson et al., 2003), and the inorganic phosphate/glycerol-3-phosphate
antiporter (GlpT) (Huang et al., 2003) were reported at resolutions of 3.5
and 3.3 A
1
4
.
6
1
3
.
9
5
.7
1
8
.
5
1
7
6
3
4
10
12
9
5 8
11
2
B
A
Figure 5.4 Proposed 3-D models of hRFC based on solved crystal structures of LacY
and GlpT and SCAM analysis, and the hypothesized substrate-binding site of hRFC.
A 3-D hypothetical model for hRFC is presented based on structure alignments
between hRFC and LacY and GlpT and fine-tuned based on experimental SCAM
data. Modeling was performed with the Modeller 8v1 auto mode (Marti-Renom
et al., 2000). All models were drawn by PyMol (DeLano, 2002). Panel A depicts a
side view of hRFC for which the extended C-terminal segment is truncated at Lys479.
TMDs 1, 2, 4, and 5 of the N-terminal region and TMDs 7, 8, 10, and 11 of the
C-terminal region are hypothesized to be involved in the formation of the hydrophilic
170 Larry H. Matherly and Zhanjun Hou
components of the putative aqueous membrane-spanning translocation
pathway flanked by TMDs 3, 6, 9, and 12. These results are consistent
with the published crystal structures of the LacY and GlpT MFS homo-
logues (Abramson et al., 2003; Huang et al., 2003). Based on data showing
the nearly complete ablation of transport activity upon cysteine substitutions
into cysteine-less hRFC, it appears that a number of amino acids are
structurally or functionally important including Arg133, Ile134, Ala135,
Tyr136, and Ser138 in TMD4; Tyr281 in TMD7; Ser313 in TMD8; and
Arg373 in TMD10. A functionally important role was also suggested for
Lys411 in TMD11 (Deng et al., 2008; Sharina et al., 2001; Witt and
Matherly, 2002). Studies with the activated NHS-MTX ester established
covalent modification at this site, thus implicating Lys411 in binding the
glutamate portion of folate substrates (Deng et al., 2008).
Notably, all these amino acids are highly conserved between species as
diverse as Xenopus, zebrafish, mice, humans, and cattle (Fig. 5.3). Compel-
ling evidence from RFC mutant studies (see above) established that Ser313
and Arg373 are functionally important and may contribute to substrate-
binding specificity (Deng et al., 2007; Hou et al., 2006; Sadlish et al., 2002b;
Sharina et al., 2001; Zhao et al., 1999). Lys411, Ser313, and Arg373 may
easily comprise a hydrophilic binding pocket for anionic folate substrates
(Hou et al., 2006) (Fig. 5.4). Reflecting its juxtaposition to both Ser313 and
Arg373 in this model, Tyr281 may also participate in substrate binding. The
finding that individual cysteine replacements of Arg133, Ile134, Ala135,
Tyr136, and Ser138 abolished transport activity is entirely consistent with
previous reports that functionally important amino acids (Ser127, Ala132,
and Arg133) were localized to TMD4 (Brigle et al., 1995; Liu and Matherly,
2001; Sharina et al., 2001; Wong et al., 1999).
The role of the highly conserved residues at positions 40, 44, 45, 46, and
48 in RFCtransport remains a paradox. Whereas mutant studies suggested a
possible functional importance for amino acids located at positions 44, 45, 46,
and 48 in hRFC (Tse et al., 1998; Wong et al., 1999; Zhao et al., 1998a,b),
only positions 40, 44, and 48 were implicated as contributing to a substrate-
binding domain by SCAM(Cao and Matherly, 2003). Similarly, in hamster
RFC, positions 41, 46, and 49 flanking TMD1 along with positions 70 and
cavity for anionic substrate binding (colored as black). TMDs 3, 6, 9, and 12 are likely
buried in the lipid bilayer and do not directly participate in substrate binding (colored as
gray). Panel B depicts a cytosolic viewof only the TMDsegments (numbered 112 as in
Fig. 5.2) of the hRFC molecule so that the order of helix packing can be easily seen.
TMD shading is the same as in Panel A. Panel C shows an enhanced view of the
hypothetical substrate-binding site, including Lys411, Ser313, Tyr281, and Arg373, as
described in the text. Other residues that may contribute to the substrate-binding
pocket are also shown and include Arg133, Ile134, Ala135, Tyr136, and Ser138. The
physical distances between a-carboxyl groups of Lys411, Ser313, Tyr281, and Arg373
are shown in Angstrom. Adapted from Hou et al. (2006).
Structure and Function of the Reduced Folate Carrier 171
71 in or flanking TMD2 were implicated in substrate binding by the ability
of RFC substrate to protect from the inhibitory effects of biotin maleimide
(Flintoff et al., 2003).
F. Mapping helix packing associations in hRFC
Characterizing conserved charged residues in or flanking TMDs by mutant
analysis can shed light on tertiary structural elements in membrane proteins
including interactions between distal domains. Interpretation is in part based
on the notion of an energetic unfavorability associated with uncompensated
charged amino acids localized within the lipid bilayer. However, should
there be a salt bridge between residues of opposite charge, charged amino
acids localized to hydrophobic environments can be substantially stabilized
(Barril et al., 1998). Salt bridges between oppositely charged residues in
separate domains can serve to orient TMDs for membrane insertion and/or
for optimal transport function (Dunten et al., 1993; Merickel et al., 1997).
For hRFC, neutralization of the positive charge on Arg133 (TMD4) by
substitution with leucine or the negative charge on Asp88 (TMD2) by
replacement with valine abolished transport activity (Liu and Matherly,
2001). However, when both mutations were present in the same construct
(i.e., Asp88Leu/Arg133Val), transport activity was restored. This suggests
that disruption of the charge-pair by replacing either Arg133 or Asp88,
individually, with a neutral amino acid results in an unstable, unpaired
charge. However, simultaneous neutralization of both charged amino
acids results in a restoration of high levels of transport activity. These results
strongly imply that Arg133 and Asp88 form a salt bridge complex that
stabilizes the association between TMDs 2 and 4 in the hRFC tertiary
structure. Other reports have also explored RFC tertiary structure at the
level of putative charge-pairs. For murine RFC, a structural or functional
interaction between Glu45 (flanks TMD1) and Lys404 (TMD11) was
suggested because the properties of the double Glu45Lys/Lys404Glu
murine RFC mutant more closely resembled the properties of the Glu45Lys
mutant than those for the Lys404Glu mutant (Zhao et al., 2003). Further,
a cross-linking analysis of hamster RFC suggested that Arg373 (in TMD10)
is in proximity to Glu394 (flanks TMD11), implying juxtaposition of these
domains and the possible formation of a charge-pair (Sadlish et al., 2002b).
Yu et al. (1995) have described a method for assessing TMD helix
packing in polytopic membrane proteins based on disulfide formation
between paired cysteine residues in purified segments of the visual pigment
rhodopsin. This method was subsequently adapted to assess helix proximities
and tilts of the bacterial MFS transporter LacY (Kaback and Wu, 1999).
For hRFC, in situ site-directed thiol cross-linking was applied to study
the proximities and tilts of neighboring transmembrane helices 2, 5, 8, and 11
(Z. Hou and L. H. Matherly, manuscript in preparation), based on their
172 Larry H. Matherly and Zhanjun Hou
proposed orientations toward the putative hRFChydrophilic cavity and their
relative proximities in 3-D models (Hou et al., 2006; Matherly et al., 2007;
Fig. 5.4). As described above, Ser313 in TMD8 was previously implicated in
the binding of antifolate substrates for RFC (Deng et al., 2007; Zhao et al.,
1999) and was identified as irreplaceable by scanning cysteine mutagenesis
(Hou et al., 2006). TMD8 abuts TMD5 and, by SCAM, both these regions
are aqueous-accessible and contribute to the substrate-binding pocket in
hRFC (Hou et al., 2006). TMD11 includes Lys411, the primary target for
covalent radioaffinity labeling with NHS-
3
H-MTX(Deng et al., 2008), and is
aqueous-accessible by SCAM (Hou et al., 2005). Finally, TMD2 flanks
TMD11 and, by homology with LacY (Abramson et al., 2003), lines the
substrate translocation pathway and includes at least one residue (Asp88) that
is essential for transport (Liu and Matherly, 2001).
In initial studies, cysteine-less hRFC was expressed as two (TMD16 and
TMD712) half molecules, each with a cysteine residue inserted at a defined
position in the TMD16 (TMDs 2 or 5) or TMD712 (TMDs 8 or 11)
portions. Altogether, 19 cysteine-substituted TMD16/TMD712 pairs
(175/311, 174/314, 172/315, 171/317, 168/318, 167/321, 164/322,
163/325, 161/326, and 160/326 in TMDs 5/8; 74/415, 74/412, 74/411,
75/408, 78/405, 81/404, 82/404, 84/404, and 85/404 in TMDs 2/11)
were selected for coexpression and cross-linking with homobifunctional
cross-linkers of different lengths [1,1-methanediyl bismethanethiosulfonate,
3 A
; o-phenylenedimaleimide, 6 A
; p-phenylenedimaleimide, 10 A
; and
1,6-bis(maleimido)hexane, 16 A
-Ca
2
exchanger. J. Biol. Chem. 274, 910917.
Nozaki, Y., Kusuhara, H., Endou, H., and Sugiyama, Y. (2004). Quantitative evaluation of
the drugdrug interactions between methotrexate and nonsteroidal anti-inflammatory
drugs in the renal uptake process based on the contribution of organic anion transporters
and reduced folate carrier. J. Pharmacol. Exp. Ther. 309, 226234.
Patterson, D., Graham, C., Cherian, C., and Matherly, L. H. (2008). A humanized mouse
model for the reduced folate carrier. Mol. Genet. Metab. 93, 95103.
Popp, C., Gorboulev, V., Muller, T. D., Gorbunov, D., Shatskaya, N., and Koepsell, H.
(2005). Amino acids critical for substrate affinity of rat organic cation transporter 1 line
the substrate binding region in a model derived from the tertiary structure of lactose
permease. Mol. Pharmacol. 67, 16001611.
Prasad, P. D., Ramamoorthy, S., Leibach, F. H., and Ganapathy, V. (1995). Molecular
cloning of the human placental folate transporter. Biochem. Biophys. Res. Commun.
206, 681687.
Price, E. M., Sams, L., Harpring, K. M., Kempton, R. J., and Freisheim, J. H. (1986).
Photoaffinity analogues of methotrexate as probes for dihydrofolate reductase structure
and function. Biochem. Pharmacol. 35, 43414343.
Russel, F. G., Masereeux, R., and van Aubel, R. A. (2002). Molecular aspects of renal
anionic drug transport. Annu. Rev. Physiol. 64(2002), 563594.
Qiu, A., Jansen, M., Sakaris, A., Min, S. H., Chattopadhyay, S., Tsai, E., Sandoval, C.,
Zhao, R., Akabas, M. H., and Goldman, I. D. (2006). Identification of an intestinal folate
transporter and the molecular basis for hereditary folate malabsorption. Cell 127,
917928.
Rong, N., Selhub, J., Goldman, B. R., and Rosenberg, I. H. (1991). Bacterially synthesized
folate in rat large intestine is incorporated into host tissue folate polyglutamates. J. Nutr.
121, 19551959.
Rothem, L., Ifergan, I., Kaufman, Y., Priest, D. G., Jansen, G., and Assaraf, Y. G. (2002).
Resistance to multiple novel antifolates is mediated via defective drug transport resulting
from clustered mutations in the reduced folate carrier gene in human leukemia cell lines.
Biochem. J. 367, 741750.
180 Larry H. Matherly and Zhanjun Hou
Rothem, L., Aronheim, A., and Assaraf, Y. G. (2003). Alterations in the expression of
transcription factors and the reduced folate carrier as a novel mechanism of antifolate
resistance in human leukemia cells. J. Biol. Chem. 278, 89358941.
Roy, Y. G., Tolner, K., Chiao, B., and Sirotnak, J. H. (1998). A single amino acid difference
within the folate transporter encoded by the murine RFC-1 gene selectively alters its
interaction with folate analogues. Implications for intrinsic antifolate resistance and
directional orientation of the transporter within the plasma membrane of tumor cells.
J. Biol. Chem. 273, 25262531.
Russel, F. G., Masereeux, R., and van Aubel, R. A. (2002). Molecular aspects of renal
anionic drug transport. Annu. Rev. Physiol. 64, 563594.
Sadlish, H., Murray, R. C., Williams, F. M., and Flintoff, W. F. (2000). Mutations in the
reduced-folate carrier affect protein localization and stability. Biochem. J. 346, 509518.
Sadlish, H., Williams, F. M., and Flintoff, W. F. (2002a). Cytoplasmic domains of the reduced
folate carrier are essential for trafficking, but not function. Biochem. J. 364, 777786.
Sadlish, H., Williams, F. M., and Flintoff, W. F. (2002b). Functional role of arginine 373 in
substrate translocation by the reduced folate carrier. J. Biol. Chem. 277, 4210542112.
Said, H. M. (2004). Recent advances in carrier mediated intestinal absorption of water
soluble vitamins. Annu. Rev. Physiol. 66, 419446.
Saier, M. H., Jr., Beatty, J. T., Goffeau, A., Harley, K. T., Heijne, W. H., Huang, S. C.,
Jack, D. L., Jahn, P. S., Lew, K., Liu, J., Pao, S. S., Paulsen, I. T., et al. (1999). The major
facilitator superfamily. J. Mol. Microbiol. Biotechnol. 1, 257279.
Salas-Burgos, A., Iserovich, P., Zunia, F., Vera, J. C., and Fischbarg, J. (2004). Predicting the
three dimensional structure of the human facilitative glucose transporter Glut1 by a novel
evolutionary homology strategy: Insights on the molecular mechanism of substrate
migration, and binding sites for glucose and inhibitory molecules. Biophys. J. 87,
29902999.
Salazar, M. D., and Ratnam, M. (2007). The folate receptor: What does it promise in tissue-
targeted therapeutics? Cancer Metastasis Rev. 26, 141152.
Schuetz, J. D., Matherly, L. H., Westin, E. H., and Goldman, I. D. (1988). Evidence for a
functional defect in the translocation of the methotrexate transport carrier in a
methotrexate-resistant murine L1210 leukemia cell line. J. Biol. Chem. 263, 98409847.
Sharina, I. G., Zhao, R., Wang, Y., Babani, S., and Goldman, I. D. (2001). Mutational
analysis of the functional role of conserved Arginine and Lysine residues in transmem-
brane domains of the murine reduced folate carrier. Mol. Pharmacol. 59, 10221028.
Sharina, I. G., Zhao, R., Wang, Y., Babani, S., and Goldman, I. D. (2002). Role of the
C-terminus and the long cytoplasmic loop in reduced folate carrier expression and
function. Biochem. Pharmacol. 63, 17171724.
Sirotnak, F. M. (1985). Obligate genetic expression in tumor cells of a fetal membrane
property mediating folate transport: Biological significance and implications for
improved therapy of human cancer. Cancer Res. 45, 39924000.
Sirotnak, F. M., and Donsbach, R. C. (1974). Stereochemical characteristics of the folate-
antifolate transport mechanism in L1210 leukemia cells. Cancer Res. 34, 371377.
Sirotnak, F. M., Kurita, S., and Hutchison, D. J. (1968). On the nature of a transport
alteration determining resistance to amethopterin in the L1210 leukemia. Cancer Res.
28, 7580.
Sirotnak, F. M., Chello, P. L., Moccio, D. M., Kisliuk, R. L., Combepine, G.,
Gaumont, Y., and Montgomery, J. A. (1979). Stereospecificity at carbon 6 of formylte-
trahydrofolate as a competitive inhibitor of transport and cytotoxicity of methotrexate
in vivo. Biochem. Pharmacol. 28, 29932997.
Sirotnak, F. M., Moccio, D. M., Kelleher, L. E., and Goutas, L. J. (1981). Relative
frequency and kinetic properties of transport-defective phenotypes among methotrexate
resistant L1210 clonal cell lines derived in vivo. Cancer Res. 41, 44424452.
Structure and Function of the Reduced Folate Carrier 181
Sirotnak, F. M., Moccio, D. M., and Yang, C. H. (1984). A novel class of genetic variants of
the L1210 cell up-regulated for folate analogue transport inward. Isolation, characteriza-
tion, and degree of metabolic instability of the system. J. Biol. Chem. 259, 1313913144.
Slotboom, D. J., Konings, W. N., and Lolkema, J. S. (2001). Cysteine-scanning mutagenesis
reveals a highly amphipathic, pore-lining membrane-spanning helix in the glutamate
transporter GltT. J. Biol. Chem. 276, 1077510781.
Spector, C., and Johanson, C. (2006). Micronutrient and urate transport in choroid plexus
and kidney: Implications for drug therapy. Pharm. Res. 23, 25152524.
Spector, R., and Lorenzo, A. V. (1975). Folate transport by the choroid plexus in vivo. Science
187, 540542.
Stokstad, E. L. R. (1990). Historical perspective on key advances in the biochemistry and
physiology of folates. In Folic Acid Metabolism in Health and Disease (M. F. Picciano,
E. L. R. Stokstad, and J. F. Gregory, eds.), pp. 121. Wiley-Liss, New York.
Suleiman, S. A., and Spector, R. (1981). Purification and characterization of a folate binding
protein from porcine choroid plexus. Arch. Biochem. Biophys. 208, 8794.
Sweiry, J. H., and Yudlievich, D. L. (1985). Transport of folates at maternal and fetal sides of
the placenta. Lack of inhibition by methotrexate. Biochim. Biophys. Acta 821, 497501.
Tse, A., Brigle, K., Taylor, S. M., and Moran, R. G. (1998). Mutations in the reduced folate
carrier gene which confer dominant resistance to 5,10-dideazatetrahydrofolate. J. Biol.
Chem. 273, 2595325960.
Wang, Y., Zhao, R., Russell, R. G., and Goldman, I. D. (2001). Localization of the murine
reduced folate carrier as assessed by immunohistochemical analysis. Biochim. Biophys. Acta
1513, 4954.
Westerhof, G. R., Schornagel, J. H., Kathmann, I., Jackman, A. L., Rosowsky, A.,
Forsch, R. A., Hynes, J. B., Boyle, F. T., Peters, G. J., Pinedo, H. M., and Jansen, G.
(1995). Carrier- and receptor-mediated transport of folate antagonists targeting folate-
dependent enzymes: Correlates of molecular-structure and biological activity. Mol.
Pharmacol. 48, 459471.
Whetstine, J. R., Flatley, R. M., and Matherly, L. H. (2002). The human reduced folate
carrier gene is ubiquitously and differentially expressed in normal human tissues: Identi-
fication of seven non-coding exons and characterization of a novel promoter. Biochem. J.
367, 629640.
White, J. C., Bailey, B. D., and Goldman, I. D. (1978). Lack of stereospecificity at carbon
6 of methyltetrahydrofolate transport in Ehrlich ascites tumor cells. J. Biol. Chem.
253, 242245.
Williams, F. M. R., and Flintoff, W. F. (1995). Isolation of a human cDNA that comple-
ments a mutant hamster cell defective in methotrexate uptake. J. Biol. Chem. 270,
29872992.
Williams, F. M. R., Murray, R. C., Underhill, T. M., and Flintoff, W. F. (1994). Isolation of
a hamster cDNA clone coding for a function involved in methotrexate uptake. J. Biol.
Chem. 269, 58105816.
Witt, T. L., and Matherly, L. H. (2002). Identification of lysine-411 in the human reduced
folate carrier as an important determinant of substrate selectivity and carrier function by
systematic site-directed mutagenesis. Biochim. Biophys. Acta 1567, 5662.
Witt, T. L., Stapels, S., and Matherly, L. H. (2004). Restoration of transport activity by co-
expression of human reduced folate carrier half molecules in transport impaired K562
cells: Localization of a substrate binding domain to transmembrane domains 712. J. Biol.
Chem. 279, 4675546763.
Wong, S. C., Proefke, S. A., Bhushan, A., and Matherly, L. H. (1995). Isolation of human
cDNAs that restore methotrexate sensitivity and reduced folate carrier activity in metho-
trexate transport- defective Chinese hamster ovary cells. J. Biol. Chem. 270,
1746817475.
182 Larry H. Matherly and Zhanjun Hou
Wong, S. C., McQuade, R., Proefke, S. A., Bhushan, A., and Matherly, L. H. (1997).
Human K562 transfectants expressing high levels of reduced folate carrier but exhibiting
low transport activity. Biochem. Pharmacol. 53, 199206.
Wong, S. C., Zhang, L., Proefke, S. A., and Matherly, L. H. (1998). Effects of the loss of
capacity for N-glycosylation on the transport activity and cellular localization of the
human reduced folate carrier. Biochim. Biophys. Acta 1375, 612.
Wong, S. C., Zhang, L., Witt, T. L., Proefke, S. A., Bhushan, A., and Matherly, L. H.
(1999). Impaired membrane transport in methotrexate-resistant CCRF-CEM cells
involves early translation termination and increased turnover of a mutant reduced folate
carrier. J. Biol. Chem. 274, 1038810394.
Yang, C.-H., Sirotnak, F. M., and Dembo, M. (1984). Interaction between anions and the
reduced folate/methotrexate transport system in L1210 cell plasma membrane vesicles:
Directional symmetry and anion specificity for differential mobility of loaded and
unloaded carrier. J. Membr. Biol. 70, 285292.
Yang, C. H., Pain, J., and Sirotnak, F. M. (1992). Alteration of folate analogue transport
inward after induced maturation of HL-60 leukemia cells. Molecular properties of the
transporter in an overproducing variant and evidence for down-regulation of its synthesis
in maturating cells. J. Biol. Chem. 267, 66286634.
Yang, R., Sowers, R., Mazza, B., Healey, J. H., Huvos, A., Grier, H., Bernstein, M.,
Beardsley, G. P., Krailo, M. D., Devidas, M., Bertino, J. R., Meyers, P., et al. (2002).
Sequence alterations in the reduced folate carrier are observed in osteosarcoma tumor
samples. Clin. Cancer Res. 9, 837844.
Yin, Y., He, X., Szewczyk, P., Nguyen, T., and Chang, G. (2006). Structure of the
multidrug transporter EmrD from Escherichia coli. Science 312, 741744.
Yu, H., Kono, M., McKee, T. D., and Oprian, D. D. (1995). A general method for mapping
tertiary contacts between amino acid residues in membrane-embedded proteins. Biochemistry
34, 1496314969.
Zeng, F. Y., Hopp, A., Soldner, A., and Wess, J. (1999). Use of a disulfide cross-linking
strategy to study muscarinic receptor structure and mechanisms of activation. J. Biol.
Chem. 274, 1662916640.
Zhang, X., Shirahatti, N. V., Mahadevan, D., and Wright, S. H. (2005). A conserved
glutamate residue in transmembrane helix 10 influences substrate specificity of rabbit
OCT2 (SLC22A2). J. Biol. Chem. 280, 3481334822.
Zhao, R., and goldman, I. D. (2003). Resistance to antifolates. Oncogene 22, 74317457.
Zhao, R., and Goldman, I. D. (2007). The molecular identity and characterization of a
proton-coupled folate transporterPCFT; biological ramifications and impact on the
activity of pemetrexed. Cancer Metastasis Rev. 26, 129139.
Zhao, R., Assaraf, Y. G., and Goldman, I. D. (1998a). A mutated murine reduced folate
carrier (RFC1) with increased affinity for folic acid, decreased affinity for methotrexate,
and an obligatory anion requirement for transport function. J. Biol. Chem. 273,
1906519071.
Zhao, R., Assaraf, Y. G., and Goldman, I. D. (1998b). A reduced carrier mutation produces
substrate-dependent alterations in carrier mobility in murine leukemia cells and metho-
trexate resistance with conservation of growth in 5-formyltetrahydrofolate. J. Biol. Chem.
273, 78737879.
Zhao, R., Gao, F., and Goldman, I. D. (1999). Discrimination among reduced folates and
methotrexate as transport substrates by a phenylalanine substitution for serine within the
predicted eighth transmembrane domain of the reduced folate carrier. Biochem. Pharmacol.
58, 16151624.
Zhao, R., Gao, F., Babani, S., and Goldman, I. D. (2000a). Sensitivity to 5,10-dideazatetra-
hydrofolate is fully conserved in a murine leukemia cell line highly resistant to
Structure and Function of the Reduced Folate Carrier 183
methotrexate due to impaired transport mediated by the reduced folate carrier. Clin.
Cancer Res. 6, 33043311.
Zhao, R., Gao, F., Liu, L., and Goldman, I. D. (2000b). The reduced folate carrier in L1210
murine leukemia cells is a 58 kDa protein. Biochim. Biophys. Acta 1466, 710.
Zhao, R., Gao, F., Wang, P. J., and Goldman, I. D. (2000c). Role of the amino acid 45
residue in reduced folate carrier function and ion-dependent transport as characterized by
site-directed mutagenesis. Mol. Pharmacol. 57, 317323.
Zhao, R., Gao, F., Wang, Y., Diaz, G. A., Gelb, B. D., and Goldman, I. D. (2001a). Impact
of the reduced folate carrier on the accumulation of active thiamin metabolites in murine
leukemia cells. J. Biol. Chem. 276, 11141118.
Zhao, R., Russell, R. G., Wang, Y., Liu, L., Gao, F., Kneitz, B., Edelman, W., and
Goldman, I. D. (2001b). Rescue of embryonic lethality in reduced folate carrier-deficient
mice by maternal folic acid supplementation reveals early neonatal failure of hemato-
poietic organs. J. Biol. Chem. 276, 1022410228.
Zhao, R., Wang, Y., Gao, F., and Goldman, I. D. (2003). Residues 45 and 404 in the
murine reduced folate carrier may interact to alter carrier binding and mobility. Biochim.
Biophys. Acta 1613, 4956.
Zhou, F., and You, G. (2007). Molecular insights into the structure-function relationship of
organic anion transporters OATs. Pharm. Res. 24, 2836.
184 Larry H. Matherly and Zhanjun Hou
C H A P T E R S I X
Renal Conservation of Folates: Role
of Folate Transport Proteins
Vijaya L. Damaraju,*
,
Carol E. Cass,*
,
and Michael B. Sawyer*
,
Contents
I. Introduction 186
II. Physicochemical Properties, Protein Binding,
and Water Solubility 186
III. Role of Folates in Genomic Stability 188
IV. Folates and the Kidney 188
V. Folate Transport Proteins 189
A. a-Folate receptor 189
B. Reduced folate carrier 190
C. Proton-coupled folate transporter 191
D. Organic anion transporters and multidrug resistance protein 192
VI. Localization of Putative Folate Transporters in Kidney 192
VII. Clinical Studies of Renal Handling of Folates and Antifolates 193
A. Folates 193
B. Antifolates 194
VIII. Role of Renal Folate Conservation in Ethanol-Related
Folate Deficiency 196
IX. Role of Folate Transport Processes in Renal Conservation
of Folates and Antifolates 197
References 198
Abstract
Folates play vital roles in one-carbon metabolism that produces the early
substrates necessary for nucleotide synthesis and salvage. Folates are essen-
tial vitamins in that humans cannot synthesize them and are totally dependent
on the diet to obtain them. As water-soluble vitamins, they would be easily
filtered by the kidney and lost to the tubular fluid but for a highly efficient renal
conservation mechanism. This renal folate trap is made up of a-folate recep-
tors and reduced folate carriers. The locations of these transporters are such
Vitamins and Hormones, Volume 79
#
2008 Elsevier Inc.
ISSN 0083-6729, DOI: 10.1016/S0083-6729(08)00406-8 All rights reserved.
* Department of Oncology, University of Alberta, Edmonton, Alberta T6G 1Z2, Canada
{
Department of Experimental Oncology, Cross Cancer Institute, Edmonton, Alberta T6G 1Z2, Canada
{
Department of Experimental Oncology, Cross Cancer Institute, Edmonton, Alberta T6G 1Z2, Canada
185
that they direct folate transport from the apical/luminal sides of kidney cells to
the basolateral/plasma sides. In addition, other transporters such as organic
anion transporters and multidrug resistance proteins are also found in kidney
cells and play a role in renal elimination of folate analogues such as antifolate
cancer chemotherapy drugs. This chapter discusses how these transporter
activities manifest themselves in folate and antifolate pharmacokinetics. It also
discusses effects of alcohol on renal reabsorption of folates. 2008 Elsevier Inc.
I. Introduction
This chapter reviews the current understanding of renal handling of
folates in mammals and the molecular processes underlying folate homeo-
stasis. Various folate analogues and their chemical structures and properties
are described briefly followed by a description of the roles that folates play in
one-carbon metabolism, especially their pivotal role in thymidine synthesis.
Then the early studies that demonstrated the kidneys role in maintaining
the bodys folate stores are reviewed.
Roles of several folate transport proteins including a-folate receptors
(aFRs), reduced folate carriers (RFCs), proton-coupled folate transporters
(PCFTs), organic anion transporters (OATs), and multidrug resistance pro-
teins (MRPs) in renal reabsorption are then reviewed. Clinical implications
of renal reabsorption of folates and antifolates are also discussed.
II. Physicochemical Properties, Protein
Binding, and Water Solubility
The term folate can have two definitions: the anion of folic acid or the
group of structurally related naturally occurring vitamins. Naturally occur-
ring folates are composed of seven members, the chemical structures of
which are presented in Fig. 6.1. Folates consist of a 2-amino-4-hydroxy-
pteridine (pterin) group conjugated by a methylene group to r-amino
benzoic acid, which in turn is linked to one or more glutamic acid residues.
The form most commonly used clinically is folic acid, also known as
pteroylglutamate (PteGlu), which is the fully oxidized form of folate and
is the most stable. The anionic and hydrophilic folate (vitamin B
9
, isolated
from spinach leaves in 1941) occurs physiologically in its reduced form (M
r
440) and belongs to the group of water-soluble B vitamins. Although
the pteridine ring can be synthesized by mammals, inability to couple it
to r-aminobenzoic acid makes folate an essential vitamin in mammals (Birn,
2006). The first step in absorption of dietary folates is hydrolysis of the
glutamate residues conjugated to the PteGlu backbone by g-glutamylhydro-
lase (glutamate carboxypeptidase II) producing the monoglutamate form.
186 Vijaya L. Damaraju et al.
Some folates are metabolized by intestinal cells by adding a methyl group to
the fifth nitrogen of the pterin ring to produce 5-methyltetrahydrofolate (5-
methylTHF). 5-MethylTHF is the most common form that is found in
plasma. Serum concentrations of the major circulating form,
5-methylTHF, range from 5 to 30 nM. Intracellular pools of reduced tetra-
hydrofolates serve as cofactors in several metabolic processes (Stover, 2004).
The vitamin is present in serum either free or bound to carrier proteins like
the folate-binding protein (Holm et al., 1980) or other serum proteins like
albumin (Soliman and Olesen, 1976). In humans, there is a wide variation in
the fraction of protein-bound folate from 20% to 65% (Zettner and Duly,
1978). Serum folates exist mostly in the monoglutamate form, which is the
form that can be readily transported across cell membranes. Once inside cells,
folates are converted to polyglutamate forms by addition of several glutamic
acid residues by folylpolyglutamate synthetase (Balinska et al., 1982). This leads
to retention of folate pools inside the cells for subsequent steps in metabolism.
Position
5,6,7,8
N
5
N
5
N
5
N
10
N
510
N
510
Metabolic step
Generation of formate
Homocysteine to methionine
Histidine metabolism
Synthesis of purines
Synthesis of purines
Synthesis of purines
Synthesis of thymidylate
Derivative
Tetrahydrofolate
5-Methyltetrahydrofolate
Formiminotetrahydrofolate
Folinic acid
10-Formyltetrahydrofolate
5,10-Methenyltetrahydrofolate
5,10-Methylenetetrahydrofolate
Substituent
-H
-CH
3
-CNH
2
-CHO
-CHO
-CH-
-CH
2
-
N
N
N
N
OH
N
H
HN
O
O
OH
O HO
H
2
N
2-Amino-6-((P-((1,3-dicarboxypropyl)carbamoyl)anilino)methyl)-4-pteridinol
(Folic acid)
5
10
N
N
N
N
N H
2
N
HN
O
O
OH
O HO
H
2
N
5
10
CH
3
4-amino-10-methyl folic acid (methotrexate)
Figure 6.1 Structures of folic acid and derivatives and their role in metabolism.
Renal Handling of Folates and Antifolates 187
III. Role of Folates in Genomic Stability
Folate is an essential vitamin that gives rise to cofactors in one-carbon
transfer reactions required for the de novo synthesis of purines and pyrimidines.
Various folates serve as essential donors and receptors for one-carbon metabo-
lism reactions. Possibly, the most important and crucial of these is methylation
of deoxyuridine monophosphate (dUMP) to deoxythymidine monopho-
sphate (dTMP). In this reaction, 5,10-methylene-H
4
PteGlu
n
donates a methyl
group to dUMP to form dTMP, a reaction that is catalyzed by thymidylate
synthase. The dihydrofolate (H
2
PteGlu
n
) produced is reduced by dihydrofo-
late reductase to tetrahydrofolate (H
4
PteGlu
n
). 5,10-Methylene tetrahydrofo-
late is regenerated by serine hydroxymethyl transferase. Folates are not only
involved in pyrimidine metabolism but 5,10-methenyl-H
4
PteGlu
n
and
10-formyl-H
4
PteGlu
n
are also involved in synthesizing the inosine monopho-
sphate that is used to synthesize adenosine monophosphate and guanosine
monophosphate. Folates therefore playa crucial role insynthesis of nucleotides.
Low folate pools result in reduced availability of nucleotides for DNA
synthesis and repair. Chronic lowfolate levels lead to uracil misincorporation
in place of thymine that results in a futile repair cycle that causes frequent
damage to DNA and chromosomes leading to malignant transformation
(Blount and Ames, 1994; Reidy, 1987). Folate also has important roles in
regulating gene expression through its role in the methylation of cytosine
residues and in DNA repair. S-adenosylmethionine is the methyl donor for
methylation of cytosines in DNA and in other cellular biochemical reactions.
IV. Folates and the Kidney
Folates are hydrophilic anionic molecules that require tightly regu-
lated processes for facilitating the transport of natural folates and antifolates.
One of the challenges that the body faces is maintaining adequate folate
stores and this is complicated by folates physiochemical characteristics that
in the absence of reabsorption mechanisms would result in folate being
freely filtered at the renal glomerulus and lost in the urine. Plasma
5-methylTHF is the major form of folate and unbound 5-methylTHF is
freely filtered at glomeruli followed by reabsorption in proximal tubules
(Goresky et al., 1963). Plasma concentrations of 5-methylTHF are 20 ng/ml
and because the kidneys glomerular filtration rate is 125 ml/min, 2000 mg
of 5-methylTHF is filtered at glomeruli each day, with only 1050 mg being
lost in urine (Birn, 2006; Birn et al., 1997). The central role of the kidney in
maintaining the 5-methylTHF pool was first suggested by Johns et al. (1961).
In their study, they administered increasing doses of radiolabeled folic acid to
188 Vijaya L. Damaraju et al.
healthy volunteers and found that at high doses, renal excretion of folic acid
approximated renal excretion of inulin, which provides a measure of renal
glomerular filtration. The same researchers confirmed the earlier study more
accurately using [
3
H] folic acid with a higher specific activity in studies
using dogs. They also demonstrated that methotrexate (MTX), a folate
analogue used in cancer treatment, displaced [
3
H] folic acid from the binding
sites for folic acid.
Mechanisms of renal folate reabsorption remained unsolved until the
studies of Selhub et al. (1987) who investigated the roles of aFRs and RFCs
in the process of renal reabsorption. Selhub et al. (1987) showed that renal
clearance was less than that of inulin, suggesting reabsorption. Renal con-
servation of folate occurs by specialized processes that depend on circulating
folate concentrations and appear to involve FRs, RFCs, and possibly the
PCFT/HCP1 systems (Bhandari et al., 1991; Kamen and Caston, 1975;
McMartin et al., 1992; Morshed et al., 1997; Qiu et al., 2006; Sikka and
McMartin, 1998). In addition to the above-mentioned transporters, OATs
and MRPs were shown to have a role in renal folate transport processes.
A concerted action of these proteins is required to provide mammalian cells
with adequate folate cofactors for nucleotide synthesis. Folate uptake pro-
cesses in apical versus basolateral membranes are important determinants of
the vectorial flow of substrates across cells. In addition to maintaining
cellular folate homeostasis, folate membrane transporters play an important
role in antitumor activity of many antifolates used in cancer therapy.
V. Folate Transport Proteins
A. a-Folate receptor
It is accepted as fact that folate homeostasis is maintained by renal reabsorp-
tion and that aFRs play a major role in this process. The aFR gene of
humans is found on chromosome 11q13 (Ragoussis et al., 1992) and
encodes a 1.28-kb mRNA (Sadasivan and Rothenberg, 1989). aFR is a
single polypeptide of 225 amino acids with a molecular mass of 26,252
(Sadasivan and Rothenberg, 1989). aFRs are anchored to apical brush
border membranes by glycosyl-phosphatidylinositol and are found in
cholesterol-rich membrane domains called caveolae (Lacey et al., 1989;
Rothberg et al., 1990). The major folate-binding protein of kidney, the
aFR, is responsible for 5-methylTHF reabsorption.
The role of the aFR, which was initially termed folate-binding pro-
tein (Selhub and Rosenberg, 1978), in folate renal handling was indicated
when it was identified in kidney proximal tubules (Kamen and Caston,
1975). Selhub and Franklin (1984) determined the proteins mass and
composition and demonstrated that it bound MTX. They also showed
Renal Handling of Folates and Antifolates 189
that the affinity of the aFR for a folate derivative was inversely related to its
clearance and, although aFR had the lowest affinity for MTX of the folate
derivatives tested, MTX was also reabsorbed. McMartin et al. (1992) studied
folate transport by human renal proximal tubule epithelial cells (RPTEC)
and demonstrated that aFRs were responsible for folate reabsorption and
showed that this reabsorption was pH dependent being maximal at pH 5.0.
Morshed et al. studied the transepithelial transport of folates using renal
proximal tubule epithelial cells grown on collagen-coated inserts
(McMartin et al., 1992; Morshed et al., 1997). They showed (using probene-
cid as an RFC inhibitor) that folate uptake at apical membranes was due to
aFRs and RFCs, whereas folate transport at basolateral membranes was due
to RFCs. aFRs were localized to brush border membranes of proximal
tubular epithelial cells in human kidney (Weitman et al., 1992). More
recently, the importance of aFRs in renal folate transport and conservation
was demonstrated in mice in which the genes folbp1 and folbp2 were deleted
(Birn et al., 2005). Elevated (100 times higher than in wild type) folate renal
clearances at low and high folate intakes were demonstrated in mice in whom
folbp1 (similar to aFRs in humans) was knocked out thus implicating folbp1 in
folate renal reabsorption. aFRs have high affinities for folic acid and MTX
with K
m
values of 1 and 100 nM, respectively ( Jansen et al., 1999). aFRs
have higher affinities for a number of folate and antifolate compounds than
bFRs (Brigle et al., 1994). Differences in binding affinities were shown to be
due to amino acid sequence differences in the binding sites between bFRs
(Leu-49, Phe-104, and Gly-164) and aFRs (Ala, Val, and Glu) (Maziarz et al.,
1999). Folate uptake by membrane bound aFRs was postulated to occur by
endocytotic processes. Folates bound to aFRs form vesicles that are then
endocytosed into cytosol, after which acidification releases folates from the
receptor complexes followed by transport into cytosol (Kamen et al., 1988;
Rothberg et al., 1990). The role of potocytosis, which is another uptake
process that has been described in cultured cells (Anderson et al., 1992), in
cellular uptake of folates is considered more controversial than the more
generally accepted endocytotic processes.
B. Reduced folate carrier
The RFC (or SLC19A1) is another membrane protein that plays a major
role in folate transport in mammalian cells with ubiquitous expression in
normal and malignant tissues, consistent with its role in tissue folate homeo-
stasis. The human RFC gene is found at position 21q22.222.3 (Moscow
et al., 1995), and has seven exons that encode a 21.4-kb mRNA
(Tolner et al., 1998). The gene encodes a 65-kDa protein that is hea-
vily glycosylated. Although there are five RFC transcript variants that
arise from alternate splicing of the 5
0
untranslated region (Moscow et al.,
1995; Nguyen et al., 1997; Prasad et al., 1995; Williams and Flintoff, 1995;
190 Vijaya L. Damaraju et al.
Wong et al., 1995), they encode the same protein. RFC exhibits broad
tissue distribution and human RFC transcripts were detected in 68 human
tissues (Whetstine et al., 2002). Immunohistochemistry studies in mice have
demonstrated RFC proteins in brush-border membranes of the jejunum,
ileum, duodenum, and colon and in basolateral membranes of renal tubular
epithelia (Wang et al., 2001). Cell-specific localization and expression of
RFC may reflect its differential roles in relation to tissue needs. RFC is a
facilitative carrier (i.e., nonconcentrative) for folates and other anionic
compounds. RFCs have different affinities for folates compared with
aFRs. Human RFCs have higher affinities for MTX and reduced folates
(K
m
15 mM) than for folic acid (K
m
100200 mM). In contrast,
aFRs have higher affinities for folic acid (K
d
1 nM) and reduced folates
(5-formyltetrahydrofolate and 5-methylTHF, K
d
1040 nM) than MTX
(K
d
150 nM) (Spinella et al., 1995). For an extensive review of RFCs, see
Matherly (2001) and Matherly and Goldman (2003).
C. Proton-coupled folate transporter
For many years, studies of uptake of folates in mammals were mainly
focused on FRs and RFCs [for recent reviews, see Matherly and
Goldman (2003)], although a low pH transport activity that differs from
both aFRs and RFC-mediated activities had been demonstrated in many
cell types. Folate transport activity with a low pH optimum was first
identified in intestinal segments, isolated cells, and membrane vesicles
(Said, 2004). The distinguishing feature of this low pH transport activity is
its similar specificity at low pH toward both oxidized and reduced folates, in
marked contrast to RFCs, which have very low affinities toward oxidized
folates (Zhao and Goldman, 2007). This low pH activity was shown to be
present in cells devoid of RFC activity, thus suggesting that it is distinct
from the RFC activity (Zhao et al., 2004, 2005). Recent cloning and
functional expression of the human gene that encodes low pH activity
demonstrated the existence of the PCFT and confirmed its role in folate
transport and absorption in intestine (Qiu et al., 2006). The PCFT gene is
found on chromosome 17 in humans and the coding region has five exons.
Prior to this discovery, this carrier was reported to be a heme carrier protein,
hence the gene has been referred as PCFT/HCP1 (Qiu et al., 2006). PCFT is
found apically in brush border membranes in duodenum and upper jejunum
(Zhao and Goldman, 2007). Alow-pHtransport activity was demonstrated in
rat kidney brush border membrane vesicles (BBMVs) (Bhandari et al., 1988).
In addition to its role in intestinal folate absorption, PCFT may have a role in
FR-mediated endocytosis (Anderson et al., 1992) in kidney, central nervous
system, and other tumor tissues. FRs present on cell membranes bind extra-
cellular folates, which are internalized in endocytotic vesicles. Folates
are released from the FRs into the cytosol upon acidification of vesicles.
Renal Handling of Folates and Antifolates 191
This release of folate from vesicles to cytosol may be mediated by PCFT
present on vesicle membranes driven by high transvesicular pH gradient
(Murphy et al., 1984; Prasad et al., 1994).
D. Organic anion transporters and multidrug
resistance protein
In addition to aFRs and RFCs, additional proteins transport antifolates in
the kidney: (1) OAT2 and OAT3 are found in basolateral membranes of
RPTECs (Matherly, 2001; Qiu et al., 2006; Sun et al., 2001; van Aubel
et al., 2002; Wang et al., 2001); (2) OAT4 is found in apical membranes of
RPTECs (Takeda et al., 2002); and (3) MRP4 (van Aubel et al., 2002) and
MRP2 (Hooijberg et al., 1999) (also known as cMOAT) are found in apical
membranes of RPTECs. In contrast to rat OAT1 (Takeuchi et al., 2001),
the human homologue hOAT1 does not transport MTX (Hirohashi et al.,
1999; Lu et al., 1999). Lu et al. expressed hOAT1 in HeLa cells and showed
that MTX was not a substrate for hOAT1 (Hirohashi et al., 1999; Lu et al.,
1999). In contrast, when Takeda et al. (2002) transfected mouse proximal
tubule cells with hOAT1, MTX transport was observed, although mice
lacking the gene encoding mouse OAT1 were not used. Evidence for
MRP-mediated transport of folates was derived from cells transfected
with cDNAs expressing the genes for MRPs (Chen et al., 2002; Hirohashi
et al., 1999; Hooijberg et al., 1999; Kusuhara et al., 1998; Zeng et al., 2000,
2001). Studies with diglutamate and higher polyglutamate derivatives of
folates demonstrated that these compounds were not substrates for these
exporters (Chen et al., 2002; Zeng et al., 2001). The major role of MRPs is
to enhance substrate export from cells and thus it may play a role in the rate
and extent of folate retention in cells by regulating intracellular concentra-
tions of monoglutamyl folates. Although MRP3 does transport MTX
(Hooijberg et al., 1999), it is not found in RPTECs, but instead is localized
to distal convoluted tubules (Scheffer et al., 2002).
VI. Localization of Putative Folate
Transporters in Kidney
Renal cells transport folates vectorially from the apical to basolateral
sides of proximal tubule cells. Resulting net fluxes and intracellular con-
centrations are a function of the activities of folate transporters in apical or
basolateral membranes of proximal tubule cells in kidney cortex. From the
integration of information obtained in different experimental systems, a
model for localization of different folate transport proteins in proximal
tubule cells is now emerging (Matherly and Goldman, 2003). aFRs are
present on brush border membranes of renal tubular epithelial cells
192 Vijaya L. Damaraju et al.
(Weitman et al., 1992). OAT-K1, a renal organic anion transporter, med-
iates MTX uptake and is localized to brush-border membranes from rat
kidneys (Masuda et al., 1997). In recent studies, Chen et al. (2002) showed
that in addition to MRP1 and MRP3, MRP2 and MRP4 also mediate
MTX efflux. MRP2 was localized to apical brush border membranes of rat
proximal tubule cells (Schaub et al., 1997). In mice, RFC1 was localized to
basolateral membranes of renal tubular epithelia (Wang et al., 2001). Human
MRP1 and MRP3 were localized to basolateral membranes of kidney
proximal tubules (Evers et al., 1996; Scheffer et al., 2002). Some anion
transporters, rOAT1 and hOAT3, were also present on basolateral mem-
branes of kidney tubules (Cha et al., 2001; Tojo et al., 1999). Interplay of
these various transport systems would favor reabsorption of folates from the
tubular lumen. Development of a cell culture model system (see Fig. 6.2)
with transporters identified thus far that play a role in folate renal transport
might give invaluable insight into roles of each transport process in renal
reabsorption versus secretion, thus enabling selective modulation of these
processes based on cellular needs.
VII. Clinical Studies of Renal Handling
of Folates and Antifolates
A. Folates
It has been well demonstrated that folic acid and other folate derivates are
renally reabsorbed in healthy volunteers who are not aggressively hydrated.
Johns et al. (1961) studied folic acid absorption kinetics using [
3
H] folic acid
Proximal tubule
cell
FR
Oat-K1/OAT4
hOAT2/3
RFC
MRP1
MRP3
MRP 2/4
Apical
Basolateral
Figure 6.2 Vectorial localization of putative folate transporters in kidney.
Renal Handling of Folates and Antifolates 193
(with relatively low activity) in 16 adult males whose age ranged from 22 to
29 years. When they administered 1 mg/kg to subjects, less than 2% of
radioactivity was recovered in the urine after 2 h. After administering
15 mg/kg, 25% of the radioactivity was recovered in the urine. At higher
doses of 150 and 1430 mg/kg, 60% to 70% of the administered radioac-
tivity was recovered in the urine. Based on these results, they concluded that
folic acid was reabsorbed by the kidney. They subsequently replicated these
experiments in vitro using dogs and [
3
H] folic acid as reported by Goresky
et al. (1963) In this later study, dogs were anaesthetized, their ureters were
catheterized, and [
3
H] folic acid was injected directly into the renal artery
along with inulin, the latter to serve as a measure of the glomerular filtration
rate. At least 25 urine samples were then collected over the ensuing 10 min.
At high doses, folic acid elimination paralleled and closely approximated
inulin filtration; whereas at low doses, folic acid secretion was substantially
less than that of inulin filtration, suggesting renal reabsorption. Interestingly
when MTX was coadministered with inulin and low dose [
3
H] folic acid,
the folic acid elimination approximated inulin filtration, suggesting that
MTX competed with folic acid to be reabsorbed. Selhub et al. (1987)
have studied other folate analogues in rats. They administered [
3
H]
folic acid, 5-methylTHF, or MTX to rats by either intravenous bolus
or continuous infusion found that urinary clearance was least for
5-methylTHF at 0.026 ml/min, followed by 0.343 ml/min for folic acid
and 1.51 ml/min for MTX. Comparing these folate urinary clearances to
the apparent urinary clearance for inulin, it was concluded that all three
folates (folic acid, 5-methylTHF, and MTX) were reabsorbed.
B. Antifolates
In contrast to studies with folates in healthy volunteers which clearly
showed reabsorption, studies in cancer patients receiving therapeutic doses
of antifolates have been less clear with many contradicting results. Calvert
et al. (1977) studied 18 cancer patients receiving MTX treatment either as
continuous intravenous infusion or as bolus. In all but 1 of the 18 patients,
renal MTX clearance was less than the measured glomerular filtration rate,
suggesting renal reabsorption. Hendel and Nyfors (1984) studied renal
clearance of low-dose MTX in 12 patients with psoriasis who were treated
with MTX with doses ranging from 7.5 to 30 mg. They found MTX had its
highest renal clearance at high initial concentrations but as plasma concen-
trations fell so did renal MTX elimination. Similarly in the phase I study of
ZD9331, a folate analogue that cannot be polyglutamated, Goh et al. (2001)
found as the dose of ZD9331 increased, the clearance increased, suggesting
renal reabsorption. As well, Sawyer et al. (2003) in a study of the oral
formulation of ZD9331 found that elevated blood urea nitrogen levels,
194 Vijaya L. Damaraju et al.
which are a marker of increased renal reabsorption, were associated with
increased risk of toxicity from ZD9331.
In contrast, when MTX was administered to patients or animals with
hydration or urine alkalinization, the kidney secreted MTX (Huang et al.,
1979; Monjanel et al., 1979). Bourke et al. (1975) studied MTX elimination
in monkeys and found that MTX elimination exceeded glomerular filtra-
tion rates. Eleven monkeys were administered MTX to achieve plasma
MTX levels ranging from 0.04 to 32 mg/m. At concentrations exceeding
1.7 mg/m, MTX elimination was approximately equal to the glomerular
filtrationrate. Incontrast whenMTXconcentrations were less than1.7 mg/ml,
elimination was threefold higher than the glomerular filtration rate. The
investigators administered probenecid, which blocks OATs, to six addi-
tional monkeys in an attempt to determine if a saturable renal excretion
process was involved. Addition of probenecid resulted in lower MTX
elimination that approximated glomerular filtration rates as assessed by
inulin clearance, suggesting saturable renal secretion. In all of these experi-
ments, the monkeys were aggressively hydrated with Ringers lactate,
a buffered lactate hydration solution. Huang et al. (1979) studied MTX
elimination in rhesus monkeys and dogs and again found that MTX elimi-
nation exceeded inulin clearance. In the study by Huang et al., the animals
were aggressively hydrated even more than the animals in the study by
Bourke et al. (1975). Liegler et al. (1969) studied effects of salicylates, para-
aminohippuric acid, and sulfisoxazole on MTX in 15 cancer patients treated
with MTX. Patients were treated with MTX to which a known quantity of
[
3
H]MTX was added. Fourteen patients were found to have normal renal
function as measured by inulin and of these, 12 had MTX clearances greater
than the inulin clearance. In patients administered high-dose para-
aminohippuric acid, the MTX clearance dropped as was the case in the
patients administered salicylates. In contrast, sulfisoxazole had little effect on
MTX clearance.
Clinical studies of renal handling of antifolates showed great contra-
dictions compared with studies of folate renal conservation. The major
difference between studies that showed renal reabsorption and studies that
showed renal secretion was that in the latter studies patients or study
animals were aggressively hydrated and their urine was alkalinized. Hydra-
tion and alkalinization were undertaken due to the belief that antifolates
can cause renal failure because of precipitation of MTX crystals in renal
tubules, leading to mechanical obstruction. No study had adequately
addressed whether antifolates could be absorbed by proximal tubule cells
via the same processes intended for folate renal conservation. Damaraju
et al. (2005) undertook studies with brush BBMVs prepared from renal
cortex to try to resolve these contradictions. BBMVs bound all anti-
folates tested, including CB3717, ZD9331, raltitrexed, aminopterin, and
MTX. Binding of antifolates was likely due to aFRs because treatment of
Renal Handling of Folates and Antifolates 195
BBMVs with phosphatidylinositol phospholipase, which cleaves glycosyl-
phosphatidylinositol linkers of aFRs, abolished binding of antifolates to
BBMVs. In contrast to naturally occurring folates that demonstrated
maximal binding to BBMVs at a pH of 6, antifolate binding to BBMVs
increased with increasing pH in many cases or stayed constant. These
results suggested that potentially all antifolates could be reabsorbed under
certain conditions, especially in the setting of dehydration or vomiting,
which would cause the urine to become alkaline. Although the alkaline
pH in many of the studies in cancer patients or laboratory animals should
have substantially increased renal reabsorption, there was net antifolate
secretion, suggesting that hydration and rapid passage of antifolates
through proximal tubules likely caused net renal secretion of antifolates
despite the presence of aFRs in proximal tubules.
VIII. Role of Renal Folate Conservation in
Ethanol-Related Folate Deficiency
Acute and chronic alcohol ingestion has been associated with folate
deficiency. Several mechanisms have been put forward as contributing
to this folate deficiency: (1) decreased dietary intake, (2) decreased folate
absorption, and (3) increased folate renal excretion because of decrea-
sed renal reabsorption (Eichner and Hillman, 1971, 1973; Halsted, 1980).
Tamura and Halsted (1983) studied folate metabolism in monkeys using
intramuscular [
3
H] folic acid and found increased excretion of folic acid in
monkeys chronically fed alcohol compared with control monkeys fed
alcohol-free diets. Russell et al. (1983) confirmed the increased renal elimi-
nation of folates during acute alcohol ingestion in five patients with chronic
alcoholism. They did their studies in a crossover fashion in which each
patient served as his/her own control. They found that acute ethanol
ingestion was associated with a 2040% increase in renal folate losses.
McMartin et al. (1986) have extensively studied the effects of ethanol on
renal folate conservation in rat models and primary cultures of human
proximal tubule cells. They demonstrated that rats fed alcohol acutely for
14 days had increased urinary excretion of folates that was maximal 8 h
after alcohol ingestion. Alcohols effect on urinary folate excretion was
similar whether the rats had been fed for 14 days, suggesting that rats
could not adapt to alcohols effects. Muldoon and McMartin (1994) showed
in a set of experiments using isolated perfused rat kidneys that infusing
alcohol into the kidney significantly increased 5-methylTHF excretion
compared to controls. Based on studies of the isolated perfused rat kidneys,
Ross and McMartin (1996) studied alcohols effects on binding of folates to
rat brush border membranes and did not find an effect of alcohol on folate
196 Vijaya L. Damaraju et al.
binding. Hamid and Kaur (2005) studied the effect of chronic alcohol
administration on binding of folates to rat renal brush border membranes
and, in contrast to the work of McMartin et al., found a decreased B
max
without any change in the affinity (K
d
). Romanoff et al. (2007) revisited the
issue of alcohols effect on folate transport using cultured human proximal
tubule cells. In their experiments, cells were exposed to alcohol in culture
media either for 1 h or for 5 days. Transport experiments were done on
membrane inserts so that alcohol effects on apical and basolateral transport
could be dissected. They found that alcohol decreased apical transport by
2025% without affecting basolateral transport. They did not find an effect
of alcohol on the B
max
or K
d
values of brush border membranes for folate.
After 5 days of exposure, they found an upregulation of both FRs and RFC
levels. In terms of the effects that alcohol has on renal conservation of folate,
there is evidence that acute ethanol exposure decreases renal reabsorption of
folate based on both in vivo and in vitro studies. The in vitro studies have not
yet produced a convincing explanation of the cause of the altered renal
reabsorption.
IX. Role of Folate Transport Processes in Renal
Conservation of Folates and Antifolates
Folates are essential vitamins that cannot be synthesized in the human
body. They are very water soluble and have modest protein binding, which
means that circulating plasma folates are subject to extensive filtration by
glomeruli in the kidneys. The amount of folate that would be lost in the
urine would be up to tenfold the average intake of folate in the North
American diet were it not for the extensive reabsorption of folates by
proximal tubule cells. In proximal tubule cells, aFRs work together with
RFCs serving as filters to prevent loss of folates into urine. The circulating
folate pool is maintained by the folate trap in the proximal tubule cells of
the kidney; this is evidenced by the fact that aFRs K
d
values for
5-methylTHF at pH 6.07.0 are 62.47.6 nM, which encompasses typical
plasma values for 5-methylTHF.
Renal conservation of folates may be a significant factor in antifolate
pharmacokinetics. Some of the contradictions of antifolate clinical studies
may be explained by the roles aFRs play in folate reabsorption in renal
proximal tubule cells. Antifolate pharmacokinetics would be more predict-
able by considering that an administered antifolate such as MTX displaces
naturally occurring folates from the circulating plasma pool and the anti-
folate is then subject to the folate trap in the kidney. Additionally, aggressive
hydration along with administration of a naturally occurring folate would
substantially decrease renal antifolate reabsorption and possibly reduce
Renal Handling of Folates and Antifolates 197
interpatient variability of antifolate pharmacokinetics. We hypothesize that
addition of folic acid inhibits the renal reabsorption of antifolates preventing
prolonged and unwanted retention of antifolates that can lead to excess
toxicity. Considering that antifolates are integrated into the folate pool and
incorporating this concept into models of antifolate pharmacokinetics may
allow better prediction of antifolate pharmacokinetics. Further studies are
clearly warranted to examine the interplay between folate conservation and
metabolism and antifolate pharmacokinetics and renal excretion.
REFERENCES
Anderson, R. G., Kamen, B. A., Rothberg, K. G., and Lacey, S. W. (1992). Potocytosis:
Sequestration and transport of small molecules by caveolae. Science 255, 410411.
Balinska, M., Nimec, Z., and Galivan, J. (1982). Characteristics of methotrexate polygluta-
mate formation in cultured hepatic cells. Arch. Biochem. Biophys. 216, 466476.
Bhandari, S. D., Joshi, S. K., and McMartin, K. E. (1988). Folate binding and transport by rat
kidney brush-border membrane vesicles. Biochim. Biophys. Acta 937, 211218.
Bhandari, S. D., Fortney, T., and McMartin, K. E. (1991). Analysis of the pH dependence of
folate binding and transport by rat kidney brush border membrane vesicles. Proc. Soc. Exp.
Biol. Med. 196, 451456.
Birn, H. (2006). The kidney in vitamin B12 and folate homeostasis: Characterization of
receptors for tubular uptake of vitamins and carrier proteins. Am. J. Physiol. Ren. Physiol.
291, F22F36.
Birn, H., Nielsen, S., and Christensen, E. I. (1997). Internalization and apical-to-basolateral
transport of folate in rat kidney proximal tubule. Am. J. Physiol. 272, F70F78.
Birn, H., Spiegelstein, O., Christensen, E. I., and Finnell, R. H. (2005). Renal tubular
reabsorption of folate mediated by folate binding protein 1. J. Am. Soc. Nephrol. 16,
608615.
Blount, B. C., and Ames, B. N. (1994). Analysis of uracil in DNA by gas chromatography-
mass spectrometry. Anal. Biochem. 219, 195200.
Bourke, R. S., Chheda, G., Bremer, A., Watanabe, O., and Tower, D. B. (1975). Inhibition
of renal tubular transport of methotrexate by probenecid. Cancer Res. 35, 110116.
Brigle, K. E., Spinella, M. J., Westin, E. H., and Goldman, I. D. (1994). Increased expression
and characterization of two distinct folate binding proteins in murine erythroleukemia
cells. Biochem. Pharmacol. 47, 337345.
Calvert, A. H., Bondy, P. K., and Harrap, K. R. (1977). Some observations on the human
pharmacology of methotrexate. Cancer Treat. Rep. 61, 16471656.
Cha, S. H., Sekine, T., Fukushima, J. I., Kanai, Y., Kobayashi, Y., Goya, T., and Endou, H.
(2001). Identification and characterization of human organic anion transporter 3 expres-
sing predominantly in the kidney. Mol. Pharmacol. 59, 12771286.
Chen, Z. S., Lee, K., Walther, S., Raftogianis, R. B., Kuwano, M., Zeng, H., and
Kruh, G. D. (2002). Analysis of methotrexate and folate transport by multidrug resistance
protein 4 (ABCC4): MRP4 is a component of the methotrexate efflux system. Cancer
Res. 62, 31443150.
Damaraju, V. L., Hamilton, K. F., Seth-Smith, M. L., Cass, C. E., and Sawyer, M. B. (2005).
Characterization of binding of folates and antifolates to brush-border membrane vesicles
isolated from human kidney. Mol. Pharmacol. 67, 453459.
Eichner, E. R., and Hillman, R. S. (1971). The evolution of anemia in alcoholic patients.
Am. J. Med. 50, 218232.
198 Vijaya L. Damaraju et al.
Eichner, E. R., and Hillman, R. S. (1973). Effect of alcohol on serum folate level. J. Clin.
Invest. 52, 584591.
Evers, R., Zaman, G. J., van Deemter, L., Jansen, H., Calafat, J., Oomen, L. C., Oude
Elferink, R. P., Borst, P., and Schinkel, A. H. (1996). Basolateral localization and export
activity of the human multidrug resistance-associated protein in polarized pig kidney
cells. J. Clin. Invest. 97, 12111218.
Goh, B. C., Ratain, M. J., Bertucci, D., Smith, R., Mani, S., Vogelzang, N. J.,
Schilsky, R. L., Hutchison, M., Smith, M., Averbuch, S., and Douglass, E. (2001).
Phase I study of ZD9331 on short daily intravenous bolus infusion for 5 days every
3 weeks with fixed dosing recommendations. J. Clin. Oncol. 19, 14761484.
Goresky, C. A., Watanabe, H., and Johns, D. G. (1963). The renal excretion of folic acid.
J. Clin. Invest. 42, 18411849.
Halsted, C. H. (1980). Folate deficiency in alcoholism. Am. J. Clin. Nutr. 33, 27362740.
Hamid, A., and Kaur, J. (2005). Kinetic characteristics of folate binding to rat renal brush
border membrane in chronic alcoholism. Mol. Cell. Biochem. 280, 219225.
Hendel, J., and Nyfors, A. (1984). Nonlinear renal elimination kinetics of methotrexate due
to saturation of renal tubular reabsorption. Eur. J. Clin. Pharmacol. 26, 121124.
Hirohashi, T., Suzuki, H., and Sugiyama, Y. (1999). Characterization of the transport
properties of cloned rat multidrug resistance-associated protein 3 (MRP3). J. Biol.
Chem. 274, 1518115185.
Holm, J., Hansen, S. I., and Lyngbye, J. (1980). High-affinity binding of folate to a protein in
serum of male subjects. Clin. Chim. Acta 100, 113119.
Hooijberg, J. H., Broxterman, H. J., Kool, M., Assaraf, Y. G., Peters, G. J., Noordhuis, P.,
Scheper, R. J., Borst, P., Pinedo, H. M., and Jansen, G. (1999). Antifolate resistance
mediated by the multidrug resistance proteins MRP1 and MRP2. Cancer Res. 59,
25322535.
Huang, K. C., Wenczak, B. A., and Liu, Y. K. (1979). Renal tubular transport of metho-
trexate in the rhesus monkey and dog. Cancer Res. 39, 48434848.
Jansen, G., Barr, H., Kathmann, I., Bunni, M. A., Priest, D. G., Noordhuis, P., Peters, G. J.,
and Assaraf, Y. G. (1999). Multiple mechanisms of resistance to polyglutamatable and
lipophilic antifolates in mammalian cells: Role of increased folylpolyglutamylation,
expanded folate pools, and intralysosomal drug sequestration. Mol. Pharmacol. 55,
761769.
Johns, D. G., Sperti, S., and Burgen, A. S. (1961). The disposal of tritiated folic acid injected
intravenously in man. Can. Med. Assoc. J. 84, 7779.
Kamen, B. A., and Caston, J. D. (1975). Identification of a folate binder in hog kidney.
J. Biol. Chem. 250, 22032205.
Kamen, B. A., Wang, M. T., Streckfuss, A. J., Peryea, X., and Anderson, R. G. (1988).
Delivery of folates to the cytoplasm of MA104 cells is mediated by a surface membrane
receptor that recycles. J. Biol. Chem. 263, 1360213609.
Kusuhara, H., Han, Y. H., Shimoda, M., Kokue, E., Suzuki, H., and Sugiyama, Y. (1998).
Reduced folate derivatives are endogenous substrates for cMOAT in rats. Am. J. Physiol.
275, G789G796.
Lacey, S. W., Sanders, J. M., Rothberg, K. G., Anderson, R. G., and Kamen, B. A. (1989).
Complementary DNA for the folate binding protein correctly predicts anchoring to the
membrane by glycosyl-phosphatidylinositol. J. Clin. Invest. 84, 715720.
Liegler, D. G., Henderson, E. S., Hahn, M. A., and Oliverio, V. T. (1969). The effect
of organic acids on renal clearance of methotrexate in man. Clin. Pharmacol. Ther.
10, 849857.
Lu, R., Chan, B. S., and Schuster, V. L. (1999). Cloning of the human kidney PAH
transporter: Narrow substrate specificity and regulation by protein kinase C. Am. J.
Physiol. 276, F295F303.
Renal Handling of Folates and Antifolates 199
Masuda, S., Saito, H., Nonoguchi, H., Tomita, K., and Inui, K. (1997). mRNA distribution
and membrane localization of the OAT-K1 organic anion transporter in rat renal tubules.
FEBS Lett. 407, 127131.
Matherly, L. H. (2001). Molecular and cellular biology of the human reduced folate carrier.
Prog. Nucleic Acid Res. Mol. Biol. 67, 131162.
Matherly, L. H., and Goldman, D. I. (2003). Membrane transport of folates. Vitam. Horm.
66, 403456.
Maziarz, K. M., Monaco, H. L., Shen, F., and Ratnam, M. (1999). Complete mapping of
divergent amino acids responsible for differential ligand binding of folate receptors alpha
and beta. J. Biol. Chem. 274, 1108611091.
McMartin, K. E., Collins, T. D., and Bairnsfather, L. (1986). Cumulative excess urinary
excretion of folate in rats after repeated ethanol treatment. J. Nutr. 116, 13161325.
McMartin, K. E., Morshed, K. M., Hazen-Martin, D. J., and Sens, D. A. (1992). Folate
transport and binding by cultured human proximal tubule cells. Am. J. Physiol. 263,
F841F848.
Monjanel, S., Rigault, J. P., Cano, J. P., Carcassonne, Y., and Favre, R. (1979). High-dose
methotrexate: Preliminary evaluation of a pharmacokinetic approach. Cancer Chemother.
Pharmacol. 3, 189196.
Morshed, K. M., Ross, D. M., and McMartin, K. E. (1997). Folate transport proteins
mediate the bidirectional transport of 5-methyltetrahydrofolate in cultured human prox-
imal tubule cells. J. Nutr. 127, 11371147.
Moscow, J. A., Gong, M., He, R., Sgagias, M. K., Dixon, K. H., Anzick, S. L.,
Meltzer, P. S., and Cowan, K. H. (1995). Isolation of a gene encoding a human reduced
folate carrier (RFC1) and analysis of its expression in transport-deficient, methotrexate-
resistant human breast cancer cells. Cancer Res. 55, 37903794.
Muldoon, R. T., and McMartin, K. E. (1994). Ethanol acutely impairs the renal conserva-
tion of 5-methyltetrahydrofolate in the isolated perfused rat kidney. Alcohol. Clin. Exp.
Res. 18, 333339.
Murphy, R. F., Powers, S., and Cantor, C. R. (1984). Endosome pH measured in single cells
by dual fluorescence flow cytometry: Rapid acidification of insulin to pH 6. J. Cell Biol.
98, 17571762.
Nguyen, T. T., Dyer, D. L., Dunning, D. D., Rubin, S. A., Grant, K. E., and Said, H. M.
(1997). Human intestinal folate transport: Cloning, expression, and distribution of
complementary RNA. Gastroenterology 112, 783791.
Prasad, P. D., Mahesh, V. B., Leibach, F. H., and Ganapathy, V. (1994). Functional coupling
between a bafilomycin A1-sensitive proton pump and a probenecid-sensitive folate
transporter in human placental choriocarcinoma cells. Biochim. Biophys. Acta 1222,
309314.
Prasad, P. D., Ramamoorthy, S., Leibach, F. H., and Ganapathy, V. (1995). Molecular
cloning of the human placental folate transporter. Biochem. Biophys. Res. Commun. 206,
681687.
Qiu, A., Jansen, M., Sakaris, A., Min, S. H., Chattopadhyay, S., Tsai, E., Sandoval, C.,
Zhao, R., Akabas, M. H., and Goldman, I. D. (2006). Identification of an intestinal folate
transporter and the molecular basis for hereditary folate malabsorption. Cell 127,
917928.
Ragoussis, J., Senger, G., Trowsdale, J., and Campbell, I. G. (1992). Genomic organization
of the human folate receptor genes on chromosome 11q13. Genomics 14, 423430.
Reidy, J. A. (1987). Folate- and deoxyuridine-sensitive chromatid breakage may result from
DNA repair during G2. Mutat. Res. 192, 217219.
Romanoff, R. L., Ross, D. M., and McMartin, K. E. (2007). Acute ethanol exposure
inhibits renal folate transport, but repeated exposure upregulates folate transport proteins
in rats and human cells. J. Nutr. 137, 12601265.
200 Vijaya L. Damaraju et al.
Ross, D. M., and McMartin, K. E. (1996). Effect of ethanol on folate binding by isolated rat
renal brush border membranes. Alcohol 13, 449454.
Rothberg, K. G., Ying, Y. S., Kamen, B. A., and Anderson, R. G. (1990). Cholesterol
controls the clustering of the glycophospholipid-anchored membrane receptor for
5-methyltetrahydrofolate. J. Cell Biol. 111, 29312938.
Russell, R. M., Rosenberg, I. H., Wilson, P. D., Iber, F. L., Oaks, E. B., Giovetti, A. C.,
Otradovec, C. L., Karwoski, P. A., and Press, A. W. (1983). Increased urinary excretion
and prolonged turnover time of folic acid during ethanol ingestion. Am. J. Clin. Nutr.
38, 6470.
Sadasivan, E., and Rothenberg, S. P. (1989). The complete amino acid sequence of a human
folate binding protein from KB cells determined from the cDNA. J. Biol. Chem.
264, 58065811.
Said, H. M. (2004). Recent advances in carrier-mediated intestinal absorption of water-
soluble vitamins. Annu. Rev. Physiol. 66, 419446.
Sawyer, M. B., Ratain, M. J., Bertucci, D., Smith, R. P., Schilsky, R. L., Vogelzang, N. J.,
Shulman, K., Douglass, E. C., and Fleming, G. F. (2003). Phase I study of an oral
formulation of ZD9331 administered daily for 28 days. J. Clin. Oncol. 21, 18591865.
Schaub, T. P., Kartenbeck, J., Konig, J., Vogel, O., Witzgall, R., Kriz, W., and Keppler, D.
(1997). Expression of the conjugate export pump encoded by the mrp2 gene in the apical
membrane of kidney proximal tubules. J. Am. Soc. Nephrol. 8, 12131221.
Scheffer, G. L., Kool, M., de Haas, M., de Vree, J. M., Pijnenborg, A. C., Bosman, D. K.,
Elferink, R. P., van der Valk, P., Borst, P., and Scheper, R. J. (2002). Tissue distribution
and induction of human multidrug resistant protein 3. Lab. Invest. 82, 193201.
Selhub, J., and Rosenberg, I. H. (1978). Demonstration of high-affinity folate binding
activity associated with the brush border membranes of rat kidney. Proc. Natl. Acad. Sci.
USA 75, 30903093.
Selhub, J., and Franklin, W. A. (1984). The folate-binding protein of rat kidney. Purifica-
tion, properties, and cellular distribution. J. Biol. Chem. 259, 66016606.
Selhub, J., Emmanouel, D., Stavropoulos, T., and Arnold, R. (1987). Renal folate absorp-
tion and the kidney folate binding protein. I. Urinary clearance studies. Am. J. Physiol.
252, F750F756.
Sikka, P. K., and McMartin, K. E. (1998). Determination of folate transport pathways in
cultured rat proximal tubule cells. Chem. Biol. Interact. 114, 1531.
Soliman, H. A., and Olesen, H. (1976). Folic acid binding by human plasma albumin. Scand.
J. Clin. Lab. Invest. 36, 299304.
Spinella, M. J., Brigle, K. E., Sierra, E. E., and Goldman, I. D. (1995). Distinguishing
between folate receptor-alpha-mediated transport and reduced folate carrier-mediated
transport in L1210 leukemia cells. J. Biol. Chem. 270, 78427849.
Stover, P. J. (2004). Physiology of folate and vitamin B12 in health and disease. Nutr. Rev.
62, S3S12; discussion S13.
Sun, W., Wu, R. R., van Poelje, P. D., and Erion, M. D. (2001). Isolation of a family of
organic anion transporters from human liver and kidney. Biochem. Biophys. Res. Commun.
283, 417422.
Takeda, M., Babu, E., Narikawa, S., and Endou, H. (2002). Interaction of human organic
anion transporters with various cephalosporin antibiotics. Eur. J. Pharmacol. 438,
137142.
Takeuchi, A., Masuda, S., Saito, H., Doi, T., and Inui, K. (2001). Role of kidney-specific
organic anion transporters in the urinary excretion of methotrexate. Kidney Int. 60,
10581068.
Tamura, T., and Halsted, C. H. (1983). Folate turnover in chronically alcoholic monkeys.
J. Lab. Clin. Med. 101, 623628.
Renal Handling of Folates and Antifolates 201
Tojo, A., Sekine, T., Nakajima, N., Hosoyamada, M., Kanai, Y., Kimura, K., and
Endou, H. (1999). Immunohistochemical localization of multispecific renal organic
anion transporter 1 in rat kidney. J. Am. Soc. Nephrol. 10, 464471.
Tolner, B., Roy, K., and Sirotnak, F. M. (1998). Structural analysis of the human RFC-1
gene encoding a folate transporter reveals multiple promoters and alternatively spliced
transcripts with 5
0
end heterogeneity. Gene 211, 331341.
van Aubel, R. A., Smeets, P. H., Peters, J. G., Bindels, R. J., and Russel, F. G. (2002). The
MRP4/ABCC4 gene encodes a novel apical organic anion transporter in human kidney
proximal tubules: Putative efflux pump for urinary cAMP and cGMP. J. Am. Soc.
Nephrol. 13, 595603.
Wang, Y., Zhao, R., Russell, R. G., and Goldman, I. D. (2001). Localization of the murine
reduced folate carrier as assessed by immunohistochemical analysis. Biochim. Biophys. Acta
1513, 4954.
Weitman, S. D., Weinberg, A. G., Coney, L. R., Zurawski, V. R., Jennings, D. S., and
Kamen, B. A. (1992). Cellular localization of the folate receptor: Potential role in drug
toxicity and folate homeostasis. Cancer Res. 52, 67086711.
Whetstine, J. R., Flatley, R. M., and Matherly, L. H. (2002). The human reduced folate
carrier gene is ubiquitously and differentially expressed in normal human tissues: Identi-
fication of seven non-coding exons and characterization of a novel promoter. Biochem. J.
367, 629640.
Williams, F. M., and Flintoff, W. F. (1995). Isolation of a human cDNA that complements a
mutant hamster cell defective in methotrexate uptake. J. Biol. Chem. 270, 29872992.
Wong, S. C., Proefke, S. A., Bhushan, A., and Matherly, L. H. (1995). Isolation of human
cDNAs that restore methotrexate sensitivity and reduced folate carrier activity in metho-
trexate transport-defective Chinese hamster ovary cells. J. Biol. Chem. 270, 1746817475.
Zeng, H., Liu, G., Rea, P. A., and Kruh, G. D. (2000). Transport of amphipathic anions by
human multidrug resistance protein 3. Cancer Res. 60, 47794784.
Zeng, H., Chen, Z. S., Belinsky, M. G., Rea, P. A., and Kruh, G. D. (2001). Transport of
methotrexate (MTX) and folates by multidrug resistance protein (MRP) 3 and MRP1:
Effect of polyglutamylation on MTX transport. Cancer Res. 61, 72257232.
Zettner, A., and Duly, P. E. (1978). The weak binding reaction between folate and human
serum proteins. Ann. Clin. Lab. Sci. 8, 5763.
Zhao, R., and Goldman, I. D. (2007). The molecular identity and characterization of a
proton-coupled folate transporterPCFT; biological ramifications and impact on the
activity of pemetrexed. Cancer Metastasis. Rev. 26, 129139.
Zhao, R., Gao, F., Hanscom, M., and Goldman, I. D. (2004). A prominent low-pH
methotrexate transport activity in human solid tumors: Contribution to the preservation
of methotrexate pharmacologic activity in HeLa cells lacking the reduced folate carrier.
Clin. Cancer Res. 10, 718727.
Zhao, R., Hanscom, M., and Goldman, I. D. (2005). The relationship between folate
transport activity at low pH and reduced folate carrier function in human Huh7
hepatoma cells. Biochim. Biophys. Acta 1715, 5764.
202 Vijaya L. Damaraju et al.
C H A P T E R S E V E N
Exploitation of the Folate Receptor
in the Management of Cancer
and Inflammatory Disease
Christopher P. Leamon* and Ann L. Jackman
Contents
I. Introduction 204
II. Aspects of the FR 205
A. FR isoforms 205
B. FR tissue expression 206
III. Exploitation of the FR for Disease Management 208
A. FR role as a biomarker 208
B. Cancer therapy 212
C. Inflammation therapy 223
D. Modulation of FR expression levels 224
IV. Future Prospects 225
References 226
Abstract
Over the last 25 years, the folate receptor (FR) has emerged as an attractive
tumor biomarker with the potential to be exploited for therapeutic purposes.
Increasing evidence suggests that this endocytosing protein can functionally
mediate the cellular uptake and retention of natural folates, certain antifolates,
and folate-drug conjugates; the consequences of the latter two events could
result in biological modulation, including (but not limited to) tumor-targeted
cytotoxicity. Because its tissue expression profile appears to be somewhat
limited to either tissues responsible for whole body retention of folates (e.g.,
kidney and placenta), or certain pathologic tissues (e.g., tumors or activated
macrophages), the FR is believed to be a useful biological target for disease
management. Indeed, recent years have been peppered with reports of novel
FR-targeted therapies, and many have demonstrated impressive in vivo potency,
particularly against tumor xenografts, without the undesirable toxicity that often
Vitamins and Hormones, Volume 79
#
2008 Elsevier Inc.
ISSN 0083-6729, DOI: 10.1016/S0083-6729(08)00407-X All rights reserved.
* Endocyte, Inc., West Lafayette, Indiana 47906
{
Institute of Cancer Research, Section of Medicine, Sutton, Surrey, SM2 5NG United Kingdom
203
accompanies nontargeted drug regimens. This chapter will provide essential
details on the properties of the FR, including where it is expressed and how it
has been successfully manipulated for therapeutic benefit. 2008 Elsevier Inc.
I. Introduction
Folic acid (FA), or vitamin B
9
, is an essential nutrient required by all
living cells for proper metabolic maintenance of one-carbon pathways as
well as for nucleotide biosynthesis (Clifford et al., 1998). This ligand displays
extremely high affinity (K
D
100 pM) for a cell surface-oriented glyco-
protein called the folate receptor (FR), which is a glycosylphosphatidylino-
sitol (GPI)-linked protein that captures its ligands from the extracellular
milieu (Luhrs and Slomiany, 1989). The FR also binds the biologically
active and reduced form of FA, namely 5-methyltetrahydrofolate, with high
affinity (Kamen and Capdevila, 1986). Immediately after binding, the plasma
membrane surrounding the FRligand complex will invaginate to form an
internal vesicle, called an endosome (see Fig. 7.1). The pH of the vesicle
lumen is somewhat lowered through the action of proton pumps that are
colocalized in the endosome membrane, and this acidification presumably
mediates a conformational change in the FR protein to release its bound
ligand to allow for cytosolic entry (Leamon and Reddy, 2004). Whereas all
eukaryotic cells also express an anionic transporter called the reduced folate
carrier (RFC) that functions in a low-affinity, high-capacity mode to deliver
folates to the cytoplasm [see Matherly et al. (2007) for an excellent review],
expression of the FR is believed to afford cells a selective growth advantage
by enabling the capture and utilization of folates, particularly under low
folate physiological conditions (Leamon and Low, 2001).
The FR is also a recognized tumor antigen (Coney et al., 1991; Ross
et al., 1994; Weitman et al., 1992a, 1994), and because of this, methods to
exploit its presence and function have been explored for possible therapeu-
tic value. For example, some laboratories are developing anti-FR antibodies
for diagnostic and therapeutic applications (Buijs et al., 1998; Konner et al.,
2006; van Zanten-Przybysz et al., 1999), whereas others are developing
high-affinity antifolates that can be delivered inside the cell via the FR to
inhibit critical proliferative functions (Gibbs et al., 2005; Henderson et al.,
2006; Jackman et al., 2004; Theti and Jackman, 2004; Theti et al., 2003).
Finally, the FA ligand itself can be covalently modified to deliver a toxic
drug payload to FR expressing cells (Leamon and Low, 2001; Leamon
and Reddy, 2004; Leamon et al., 2007a; Reddy et al., 2005).
The purpose of this chapter is to provide the reader with a basic
understanding of FR biology, what tissues express it, and how this protein
can be exploited for the uptake and/or delivery of active pharmaceutical
204 Christopher P. Leamon and Ann L. Jackman
agents. Our reviewwill not exhaustively cover this ever expanding field. For
example, topics such as folate-targeted liposomes, nanoparticles, antisense/
gene formulations, or antibody-mediated therapies will not be covered in
this volume. Instead, we intend to focus only on small molecule-based
applications that are currently being evaluated in the clinic, or those that
are at the preclinical stage of development.
II. Aspects of the FR
A. FR isoforms
Several isoforms of the FR have been reported, and they appear to have
distinct expression patterns in normal and pathological tissues as well as
different properties. The two FR isoforms most relevant to therapeutic
Drug
Spacer
Folate
FR
H
+
H
+
H
+
Endosome
Membrane
invagination
Nucleus
Recycling
endosome
RFC
Antifolate
Late endosome
(segregation)
Figure 7.1 Schematic representation of the trafficking of folates, antifolates, and
FA-conjugates by the FR-mediated endocytosis pathway. Exogenously added ligands
bind specifically to the FR protein with high affinity. The plasma membrane invaginates
around the ligandFR complex to form an intracellular vesicle (early endosome). As
the lumen of the maturing endosome acidifies, the receptor changes conformation and
releases the ligand. Eventually, the fates of released ligands, drug cargo, and the FR are
determined during a sorting process within late endosomal elements near the peri-
nuclear region. The reduced folate carrier (RFC), which unlike the FR is an anion
transporter, can shuttle folate molecules and certain antifolates inside the cell. FAdrug
conjugates, however, are not substrates for the RFC.
Folate Receptor-Based Therapies 205
exploitation are FR-a and FR-b, both of which possess a GPI anchor and
both are functional for high-affinity FA binding, although they can display
some differences with respect to their affinity for reduced folate cofactors or
antifolates (Salazar and Ratnam, 2007). FR-g is an isoform that can be
detected in some hematopoeitic cells; however, because it is not anchored
to the plasma membrane, this protein is constitutively excreted. For this
reason, FR-g will not be discussed further. Following capture of folate
molecules, membrane-associated FR-a and FR-b appear within endocytic
compartments before they are recycled back to the cell surface. Studies
reviewed by Salazar and Ratnam (2007) suggest that there may be cell type
differences for receptor recycling kinetics that may impact the delivery
efficiency of folate or folate conjugates fromthe cell surface to the cytoplasm.
B. FR tissue expression
1. Methods
There are a number of relevant reports, some with small sample numbers,
that describe the tissue distribution of FRs. Five broad methods have been
used to compare FR expression levels. These are (1) FR mRNA levels
measured by PCR or in situ hybridization, (2) FR-a protein levels measured
by immunohistochemistry (IHC) using mono- and polyclonal antibodies,
(3) functional FR levels as measured by
3
H-FA binding to solubilized
tissue membrane preparations, (4) assessing FR tissue expression in real time
using radiodiagnostic imaging, and (5) quantifying FR expression ex vivo by
flow cytometric methods. Importantly, each method has its own particular
merits and pitfalls. For example, FR expression has been reported to be
controlled at the translational level, at least partly in some cells, thus it might
be expected that mRNA and protein levels may not always correlate. IHCis
less amenable to precise quantitation, particularly when comparing the
results from different studies, but it has the advantage in that FR distribution
within a tissue can be visualized; the latter is probably essential when
considering therapeutic exploitation of FR expression because in normal
polarized tissues, expression of this protein is restricted to the apical (luminal)
membrane surface rather than the blood-accessible, basolateral surface. Total
tissue receptor measurement using
3
H-FA has the advantage of measuring
FR that is functional for ligand binding, but this method may be less useful
for (1) polarized normal tissues, (2) assessing the degree of homogeneity
of receptor expression within a tumor, or (3) distinguishing between differ-
ent receptor isoforms. Imaging tissues for FR expression after injection
of an animal or human with a radioactive high-affinity folate-conjugate
has the advantage of measuring FR that is accessible within the blood
stream. Therefore, this latter approach is of high therapeutic relevance
(see Section III.A.2 below); but, it unfortunately cannot quantitate absol-
ute levels of FR protein. Finally, the FR can also be measured by flow
206 Christopher P. Leamon and Ann L. Jackman
cytometry on either disaggregated tumor material (Toffoli et al., 1998) or
surface of epithelial tumor cells in effusions, such as ascitic fluid (Forster
et al., 2007). One disadvantage of this method, though, is that the tissue to
be analyzed requires isolation and ex vivo manipulation.
2. Normal tissue expression
FR-a distribution in normal adult tissue is restricted to the apical membrane
surface of some polarized epithelial cells, including lung, choroid plexus,
and some glandular tissue (Weitman et al., 1992a,b). Expression is also high
in placental trophoblasts and on the luminal surface of proximal tubule
kidney epithelial cells, the latter probably being important for the reabsorp-
tion of folates from the urine (Birn et al., 1993, 1997; Garin-Chesa et al.,
1993; Holm et al., 1992; Prasad et al., 1994; Rettig et al., 1985; Weitman
et al., 1992b). FR-b is expressed on some normal hematopoeitic cells of
myelomonocytic lineage. Interestingly, this isoform is functional when
expressed on activated monocytes and macrophages, but it is nonfunctional
(unable to bind folate) when present on mature neutrophils or CD34stem
cells (Nakashima-Matsushita et al., 1999; Paulos et al., 2004; Reddy et al.,
1999; Ross et al., 1999; Turk et al., 2004).
3. Cancer expression
Elevated expression of the FR-a occurs in several cancer types, although the
actual frequency and degree of expression reported is variable, in part due to
the different methods used and the small sample numbers in some studies.
Nonmucinous ovarian cancer (the majority of ovarian cancers) was the tumor
type first to be associated with overexpression (Campbell et al., 1991).
Initially, this was in the form of reports that a monoclonal antibody (mAb;
MOv18) raised against a membrane preparation from an ovarian tumor,
detected a glycoprotein present in most ovarian tumors (Miotti et al., 1987;
Veggian et al., 1989); furthermore, it was later shown that this antigen was a
high-affinity folate-binding protein (FBP) with an amino acid sequence
identical to FBPs found in placenta and KB (human nasopharyngeal carci-
noma) tumor cells (Campbell et al., 1991; Coney et al., 1991). Several studies
confirmed that 8090% of these tumors overexpress FR-a (Parker et al.,
2005; Toffoli et al., 1997; Wu et al., 1999). Other gynecological cancers (e.g.,
50% of serous uterine tumors) also overexpress the receptor (Allard et al.,
2007; Dainty et al., 2007; Marziarz et al., 1999; Wu et al., 1999). Although the
endometrioid histological subtype may express the receptor less frequently, it
is by far the most common form of uterine cancer and therefore a consider-
able number of uterine cancer patients may benefit from some form of
FR-targeted therapy. Other tumors reported to overexpress FR-a to varying
frequencies include pediatric ependymal brain tumors, breast, colon, renal
and lung tumors, and mesothelioma (Parker et al., 2005).
Folate Receptor-Based Therapies 207
Current knowledge also suggests that functional FR-b overexpression in
cancer is restricted to myeloid leukemia and perhaps head and neck carci-
noma (Ross et al., 1994, 1999). Taken together, the total number of tumors
with FR overexpression is very large; thus, FR-targeted strategies could
have significant impact on cancer treatment for patients diagnosed with FR
overexpression.
III. Exploitation of the FR for
Disease Management
A. FR role as a biomarker
As discussed above, the FR is expressed on the apical membrane of polar-
ized epithelia in a limited number of normal tissues; however, many
different types of malignant tissues have also been found to express this
protein, and in large quantities (Coney et al., 1991; Parker et al., 2005; Ross
et al., 1994; Weitman et al., 1992a, 1994). Importantly, not all human
cancers within a particular indication will express the FR; ovarian cancer
is close at 90%, but the total positives of most other cancer types is
lower. For example, 64% (102 of 160) of renal cell carcinoma specimens
are positive, and that number decreases to less than 20% for primary
colorectal carcinoma (unpublished data, Endocyte, Inc.). Because novel
FR-targeted therapies are now being tested clinically (Amato et al., 2005;
Konner et al., 2006; Leamon et al., 2007a; Messmann et al., 2007; Reddy
et al., 2007a; Sausville et al., 2007), the FR may be considered a useful tissue-
based biomarker. Therefore, having the ability to screen patients for
FR-positive disease could certainly increase the efficiency of and decrease
the time for clinical investigations of FR-targeted therapies.
At present, two principal methods have been utilized for assessing
a patients FR status. These include an invasive tissue-based immunohisto-
chemical assay, and a noninvasive radiodiagnostic approach. The merits of
both techniques will be discussed below.
1. Immunohistochemistry
IHC has proven to be beneficial for the selection of patients that are more
likely to respond to a given therapy. Indeed, the HercepTest
IHC screen,
which is used for selecting patients with Her-2 expressing breast cancer
( Jacobs et al., 1999), enables the clinician to predict with reasonable cer-
tainly which patients may not benefit from Herceptin therapy. A related test
has also been developed for the immunohistochemical detection of the FR.
Thus, polyclonal rabbit antiserum was produced using bovine milk soluble
folate-binding protein as the antigen. Due to the high degree of homology
between the human FR and folate-binding proteins from various sources,
208 Christopher P. Leamon and Ann L. Jackman
the affinity-purified IgG fraction from this antiserum (herein called PU17)
adequately detects the presence of FR in formalin-fixed, paraffin-embedded
tissues (Endocyte, Inc.). PU17 was selected as the probe to validate the first
FR IHC method to be clinically used. Importantly, the scoring methodol-
ogy is based on a 0 to 3 scale (similar to HercepTest
100
75
50
25
0
10
0.1
A431
A431-FBP + FA
A431-FBP
KB + FA
KB
1
0.01
G
r
o
w
t
h
i
n
h
i
b
i
t
o
r
y
I
C
5
0
(
M
)
0.001
0.0001
C
B
3
7
1
7
R
a
l
t
i
t
r
e
x
e
d
P
e
m
e
t
r
e
x
e
d
P
l
e
v
i
t
r
e
x
e
d
B
G
C
9
4
5
BGC945 concentration ( M)
C
e
l
l
g
r
o
w
t
h
(
%
c
o
n
t
r
o
l
)
0.0001 0.001 0.01 0.1 1 10 100
A431
A431-FBP
) of with a
40-fold excess of free folate (). A control cohort (
, and added
to the cell. The cells were passaged at 1:10 dilution into fresh growth medium
24 h after transfection. Zeocin, 300 mg/ml was added for stable selection.
C. Western blotting procedures
For total cell extracts, cells were washed twice with PBS, scraped, and cen-
trifuged at 1000 rpm for 5 min. The pellet was suspended in denaturation
buffer (10 mmol/liter Tris, 150 mmol/liter NaCl, 2 mmol/liter EDTA,
pH 7.4, and 0.2% SDS and were denatured at 100
C for 5 min. Protein
concentration was determined using the bicinchoninic acid protein assay
and proteins were resolved on a 15% SDS acrylamide gel and transferred to a
nitrocellulose membrane. FRa was detected by incubating the membrane
238 Chheng-Orn Evans et al.
with polyclonal antihuman FRa antibody, followed by HRP-conjugated
anti-rabbit IgG secondary antibody. Visualization was carried out on X-ray
film by enhanced chemiluminescence. b-Actin expression was detected to
ensure equal protein loading.
D. Cell proliferation assay
Cells growing in log phase were lifted using EDTA and seeded in 24-well
plates (1 10
4
cells/well in a final volume of 500 ml) in triplicates and
incubated at 37
C in 5% CO
2
. The cells were incubated for 10 days and the
media were changed every 3 days. Every other day, the cells were counted
with Trypan blue in hematocytometer. Each experiment was repeated three
times.
E. Soft agar colony assay
Cells (5 10
3
) from each cell line collected using EDTA were suspended in
0.3% agar in DMEM containing 10% FBS and plated on 0.7% solidified agar
in 35 mm dishes. After 15 days of culture, colonies with more than 20 cells
were counted and photographed.
F. BrdU incorporation assay
BrdU is only incorporated into the DNA of proliferating cells. BrdU
incorporation assay was carried by pulse labeling the cells with BrdU for
30 min. The levels of BrdU incorporated into cellular DNA were quanti-
fied by anti-BrdU antibody; total DNA was stained by 7-AAD. The samples
were analyzed by BD FACScan and data were evaluated using Flowjo
software. This two-color flow cytometric analysis permits the enumeration
and characterization of cells that are actively synthesizing DNA (BrdU
incorporation) in terms of their cell cycle position (i.e., G0/1-, S-, or
G2/M-phases, defined by 7-AAD staining intensities).
G. Flow cytometric assessment of PCNA expression
PCNA is a nuclear antigen that is only expressed in proliferating cells and
absent in resting cells. For PCNA analysis, cells were washed with 1% FBS/
PBS and fixed with cold 70% ethanol overnight in 50 ml tube. After
removing ethanol, the cells were incubated with an anti-PCNA,
fluorescein-conjugated monoclonal antibody, or isotype-matched control
IgG (Pharmingen, Becton Dickinson, Jan Hose, California) for 30 min at
room temperature. The stained cells were then washed and fixed with
formaldehyde. Fluorescence data from 10
4
cells were collected and histogram
analysis was performed with the Flowjo software.
Folate Receptor Expression in Pituitary Adenomas 239
H. Folic acid binding
[
3
H]folic acid purification and solubilized folic acid binding were performed
as described (Doucette and Stevens, 2001). Cells were grown in T-25 in
folate-free RPMI for 3 days and were selected with 300 mg/ml of Zeocin
when applicable. After washing twice with PBS and centrifugation for
5 min at 1000 rpm, the pellet was solubilized in 1-ml ice-cold solubilization
buffer (25 mmol/liter TrisHCl, pH 7.5, 150 mmol/liter NaCl, 5 mmol/
liter EDTA, and 1% (v/v) Triton X-100) for 20 min. [
3
H]folic acid
(20175 nmol/liter, depending on the cell line) was added and the samples
were incubated at 37
C for 1 h in shaking water bath to allow ligand
binding to the receptor. The samples were then placed on ice for 5 min,
and 1 ml of dextran-coated charcoal solution (25 mmol/liter TrisHCl, pH
7.5, 150 mmol/liter NaCl, 5 mmol/liter EDTA, 8 mol/liter dextran, and
80 mol/liter activated charcoal) was added to each tube to absorb unbound
radiolabel. The samples were vortexed and incubated on ice for 15 min,
followed by centrifugation for 30 min at 3400 rpm at 4
C. Aliquot of 1 ml
of each supernatant was counted by scintillation counting. Nonspecific
binding was determined in each experiment by measuring [
3
H]folic
acid binding in the presence of 750-fold excess unlabeled folic acid. Specific
binding values were determined by subtracting the nonspecific value from
the respective total radioactivity.
I. Quantitative RT-qPCR
Expression of the selected genes was quantified using RT-qPCR analysis.
Briefly, total RNA was isolated using the Trizol reagent (Invitrogen) and
was reverse transcribed using Supersript II RT (Invitrogen). Specific pri-
mers were designed using Primer3 software (MIT Whitehead Institute,
http://frodo.wi.mit.edu/cgi-bin/primer3/primer3_www.cgi) and listed in
Table 8.2. All PCR reactions were performed in GeneAmp 5700 sequence
detection system (Applied Biosystems) and repeated in three separate
experiments, each sample was run in duplicate. b-Actin and b-tubulin
were used as internal control. Cycle threshold (Ct) values were obtained
and the relative differences in expression between groups were determined
using the 2(ddCT) formula, where ddCT (Ct gene of interest Ct
actin or tubulin in experimental sample) (Ct gene of interest Ct actin or
tubulin in untransfected, untreated aT 3-1 cells sample).
J. Statistical analyses
Results are expressed as mean SEM. Differences were assessed by one-
way Anova to all results except the binding assay where student t-test was
used. p < 0.05 was considered to be statistically significant.
240 Chheng-Orn Evans et al.
Table 8.2 Primer sequences used for quantitive RT-qPCR
Gene Forward primer Reverse primer
FRa GCATTTCATCCAGGACACCT GGTGTAGGAGGTGCGACAAT
NOTCH3 TGAGTGTCCAGCTGGCTATG CACAGGTGCCATTGTGTAGG
Hes-1 AAACGAAAATGCCAGCTGAT ATGCCGGGAGCTATCTTTCT
DLK1 TGTCAATGGAGTCTGCAAGG AGGGAGAACCATTGATCACG
Pit1 CACGGCTCAGAATTCAGTCA CTGATGGTTGTCCTCCGTTT
SFRP1 CGAGTTTGCACTGAGGATGA GCAGGTACTGGCTCTTCACC
ASCL-1 CATCTCCCCCAACTACTCCA CAAAGTCCATTCCCAGGAGA
b-Catenin GTGCAATTCCTGAGCTGACA CTTAAAGATGGCCAGCAAGC
IDH1 AGGTTCTGTGGTGGAGATGC GACGCCCACGTTGTATTTCT
PITX2 CTGGAAGCCACTTTCCAGAG GTACGAATAGCCGGGGTACA
RB1 GCAGTCCAAGGATGGAGAAG ACAGGGCAAGGGAGGTAGAT
TLE2 CCCTTTCACCTCATCCTTCA CTGTGCCTACCAGAGCATCA
Cyclin D1 AGTGCGTGCAGAAGGAGATT CACAACTTCTCGGCAGTCAA
ERa TTACGAAGTGGGCATGATGA ATAGATCATGGGCGGTTCAG
ERb GAAGCTGGCTGACAAGGAAC AACGAGGTCTGGAGCAAAGA
GATA3 CTGGAGGAGGAACGCTAATG CAGGGATGACATGTGTCTGG
EGFR ACACTGCTGGTGTTGCTGAC CCAAGGACCACTTCACAGT
PTTG1 GGCATCTAAGGATGGGTTGA TTCGGCAACTCTGTTGACTG
FGFR1 ATGGTTGACCGTTCTGGAAG GGAAGTCGCTCTTCTTGGTG
Actin
a
TATGCCAACACAGTGCTGTCTGG TACTCCTGCTTGCTGATCCACAT
Tubulin
b
GGAGAGCTGTGATTGCCTGC CCACCCAGTGAGTGGGTCAG
a
Wang et al. (2007).
b
Gao et al. (2004).
III. Result
In a previous study (Evans et al., 2001), we used cDNA microarray
analysis and RT-qPCR to compare expression profiles of 7075 genes in the
normal pituitary with that in different adenomas, including NF, PRL-
producing adenomas, GH-producing adenomas, and ACTH-secreting ade-
nomas. In those experiments, we found that the FRa gene was significantly
overexpressed in NF adenomas. Next, we characterized the expression of
FRa in NF and NF, PRL, GH, and ACTH pituitary adenomas (Evans
et al., 2003). The identification of FRa may further elucidate the pathways
of pituitary oncogenesis and provide effective therapeutic treatment to
pituitary tumors.
We show some of the importance results of our previous study (Evans
et al., 2003) in following sections.
A. Tumor classification
The clinical and pathological characteristics of the 39 adenomas used in the
study are listed in Table 8.3. Ten of the NF adenomas were not positive
with anterior pituitary hormone histochemistry and were designated immu-
nohistochemically negative (NF) tumors. Thirteen NF tumors stained
with one or more anterior pituitary hormones and were designated immu-
nohistochemically positive (NF). Six were classified as PRL, five as GH,
and five as ACTH-positive adenomas. With the exception of two with
cavernous sinus invasion, all tumors were noninvasive as defined by histo-
logical and radiological criteria. Other clinical features related to the tumor
are noted in Table 8.3. Five normal pituitary controls were obtained from
the National Hormone and Pituitary Program, National Institute of Diabe-
tes and Digestive and Kidney Diseases.
B. FRa mRNA expression in NF adenomas
We investigated the relative expression levels of FRa mRNA in NF
adenomas by RT-qPCR. We found that there was significant overexpres-
sion of FRa mRNA in 13 NF adenomas from 17- to 174-fold compared
with the controls ( p 0.001) (Fig. 8.1).
C. Expression of FRa protein
Whether the high levels of FRa mRNA resulted in overexpression of FRa
protein was investigated with Western blotting of 39 pituitary tumors and 5
normal pituitaries using a highly specific polyclonal antibody against this
242 Chheng-Orn Evans et al.
Table 8.3 Clinical and pathological characteristics of adenomas from patients used
in the study (Evans et al., 2003)
Patient ID Sex, Age Clinical features/tumor size IHC
93 M, 57 NF, visual loss, cavernous
sinus invasion, 1.5 cm
Neg
a
74 M, 67 NF, hypogonadism,
2.5 cm
Neg
164 M, 35 NF, visual loss, 4 cm Neg
153 M, 72 NF, hypogonadism, 2 cm Neg
155 M, 66 NF, headache, visual loss,
hypocortisolism,
1.5 cm
Neg
104 F, 61 NF, visual loss, 1.5 cm Neg
105 F, 48 NF Neg
75 M, 60 NF, visual loss, 2 cm Neg
68 M, 74 NF Neg
191 F, 44 NF Neg
65 F, 54 NF FSH 1, LH 2
77 M, 67 NF FSH 2
174 F, 53 NF FSH 1, LH 1
143 M, 60 NF FSH 1, LH 1,
ACTH 1
89 M, 62 NF FSH 2
91 M, 56 NF FSH 2, TSH 1
100 F, 65 NF FSH 1, LH 1,
ACTH 2,
PRL, GH1
112 F, 58 NF FSH 3, PRL 2,
LH 2
69 M, 67 NF FSH 2, LH 3
60 F, 90 NF FSH 1,
LH 23
198 M, 58 NF FSH 1, LH 1
208 F, 47 NF LH1
138 M, 60 NF FSH 12, LH
12
183 M, 36 Hyperprolactinemia,
visual loss, 2 cm
PRL 3
192 M, 41 Hyperprolactinemia,
visual loss, 3 cm
PRL 3
151 M, 52 Hyperprolactinemia PRL 3, GH 1
240 M, 39 Hyperprolactinemia PRL 3
244 F, 40 Hyperprolactinemia PRL 3
(continued)
Folate Receptor Expression in Pituitary Adenomas 243
Table 8.3 (continued)
Patient ID Sex, Age Clinical features/tumor size IHC
213 M, 24 Hyperprolactinemia PRL 3
123 M, 30 Acromegaly, 2 cm GH 3, TSH 3,
FSH 3, LH 3
196 F, 69 Acromegaly, 2 cm ND
b
168 F, 39 Acromegaly, 1 cm GH 3, PRL 23
218 F, 33 Acromegaly GH 3, PRL 23
145 M, 34 Acromegaly GH 3, PRL
23, TSH 1
232 F, 39 Cushings disease ACTH 2
137 F, 41 Cushings disease ND
239 F, 55 Cushings disease, cav
sinus invasion
ACTH 4
126 F, 29 Cushings disease, 1.6 cm ACTH 3
233 M, 11 Cushings disease ACTH 3,
FSH 1
a
Neg, immunostaining for GH, ACTH, PRL, FSH, LH, and TSH were negative.
b
ND, not determined.
Sex, age of patient, and a brief description of tumor type are given.
Con
0
40
80
120
160
200
FR expression in NF immunohistochemically
positive by RT-qPCR
NF immunohistochemically positive samples
F
R
m
e
s
s
a
g
e
(
r
e
l
a
t
i
v
e
t
o
c
o
n
t
r
o
l
)
77 65 174 143 89 91 100 112 69 60 198 208 138
Figure 8.1 Relative expression levels of FRa mRNA in NF adenomas by RT-
qPCR. Expression levels of each tumor were normalized to the 18S RNA of the
same sample. Fold difference was the ratio of the normalized value of each sample to
controls as described in Methods and Materials (Evans et al., 2003). Con, five normal
controls; 65138, NF adenomas. There was a significant overexpression of FRa
mRNA in 13 NF adenomas from 17- to 174-fold compared with the controls ( p
0.001). Notably, sample 208 exhibited the highest FRa mRNA expression among the
13 samples (174-fold).
244 Chheng-Orn Evans et al.
receptor. This antibody was produced by rabbits immunized with a recom-
binant FRa-glutathione S-transferase fusion protein and has been shown to
recognize the same protein as the Mab LK26 (data not shown). The results
for two normal pituitaries (N1 and N3), one NF adenoma (#65), and
9 NFadenomas are shown in Fig. 8.2A and for the 13 NFadenomas are
shown in Fig. 8.2B. Two PRL-secreting adenomas (PRL #183 and 192)
and three GH-secreting adenomas (GH #123, 196, and 168) are shown
in Fig. 8.2C, whereas four ACTH-secreting (ACTH #232, 137, 239,
and 126) adenomas are shown in Fig. 8.2D. Adenoma #65, which was
NF, was included in all of the blots to provide a basis of comparison for
the different gels. These results show that the majority (7 of 10 tumors
showing FRa expression) of the NF tumors and all of the NF tumors
overexpress FRa relative to the normal pituitaries and functional adenomas.
The results of the Western blot analysis are summarized in a box plot
shown in Fig. 8.3. The horizontal line in each box represents the median
value of FRa expression of each adenoma group and controls. The box
represents the 25th and 75th percentile range of scores. The 10 NFsamples
showed an average increase of 19-fold compared with controls, whereas the
N1
77 138 174 143 89
65 183 192 123 196 168
91 100 112 69 60 198 208 65
kDa
40.0
A
B
C D
28.8
kDa
40.0
28.8
kDa
40.0
28.8
65 232 137 239 126
kDa
40.0
28.8
N3 65 68 93 74 164 153 155 104 105 75
Figure 8.2 (AD) Western blot of FRa expression in pituitary adenomas. Total protein
(10 mg) of each sample was separated by a 15% SDS-PAGE. Immunodetection was
carried out using a polyclonal antibody, rabbit antihuman FRa IgG, as described
in Methods and Materials. Panel (A) showed FRa expression in two normal pituitaries,
one NF (65), and nine NF adenomas. Panel (B) showed FRa expression in 13 NF
adenomas. Panel (C) showed FRa expression in one NF (65), two PRL (183 and 192),
and three GH-secreting adenomas (123, 196, 168), whereas panel (D) showed one NF
(65) and four ACTH-secreting adenomas (232126).
Folate Receptor Expression in Pituitary Adenomas 245
13 NFshowed a mean increase of 36-fold relative to controls. Six PRL, five
GH, and five ACTH-secreting adenomas showed no increase compared with
controls. The KruskalWallis test on ranks of the median values of FRa
protein showed that there were significant differences between tumor groups
(w
2
29.639, degrees of freedom 5, p < 0.05). The Mann-Whitney test
showed there was a significant difference between NF compared with
controls ( p 0.014), PRL ( p 0.009), GH ( p 0.014), and ACTH-
secreting adenomas ( p 0.007). Furthermore, there was a significant differ-
ence between NF compared with controls ( p 0.001), NF ( p 0.041),
PRL ( p 0.001), GH ( p 0.001), and ACTH-secreting adenomas ( p
0.001). However, there was no significant difference between PRL, GH, and
ACTH-secreting adenomas compared with controls.
D. Assessment of FRa-binding capacity
The Western blot analysis indicates the relative amount of a protein in the
tissue or cell but does not provide any information about its functional
potential. To determine whether the overexpressed FRa protein in the
FR protein expression in pituitary adenomas
*
**
Tumor type
5
Control
0
10
10
F
o
l
d
i
n
c
r
e
a
s
e
c
o
m
p
a
r
e
d
t
o
c
o
n
t
r
o
l
(
O
D
)
20
30
40
10
NF
13
NF+
6
PRL
5
GH
5
ACTH
n =
Figure 8.3 Box plots representing the FRa protein expression by adenoma subtypes,
by Western blot. A horizontal line in each box represents the median value of FRa
protein expression of each group. Box, the 25th and 75th percentile range of scores.
Whiskers, the highest and lowest values. Numbers of pituitaries tested in each group
were n 5 for controls, 10 for NF, 13 for NF, 6 for PRL, 5 for GH, and 5 for
ACTH-secreting adenomas. *, in NFadenomas, FRa was significantly overexpressed
compared with controls ( p 0.014), PRL ( p 0.009), GH ( p 0.014), and ACTH-
secreting adenomas ( p 0.007). **, in NF adenomas, FRa was significantly over-
expressed compared with controls ( p 0.001), NF ( p 0.041), PRL ( p 0.001),
GH ( p 0.001), and ACTH-secreting adenomas ( p 0.001).
246 Chheng-Orn Evans et al.
adenomas was properly folded and had the potential to transport folates,
specific binding of folic acid was measured in the various tumors for which
sufficient tissue was available (37 pituitary tumors and 5 normal pituitaries).
In the five normal pituitary controls, folic acid binding was 0.94.8 pmol/
mg protein, with a mean of 2.2 pmol/mg protein (Fig. 8.4). In the 10 NF
samples, binding ranged from 1.6 to 136.1 pmol/mg protein, which was
0.7- to 62-fold greater than the mean of the controls. The 13 NF
adenomas bound between 7.6 and 242.7 pmol/mg protein, which was 3-
to 110-fold higher than the mean of control samples. Folic acid binding was
very low (0.022.1 pmol/mg protein) in five PRL, five GH, and four
ACTH-secreting adenomas.
E. Immunohistochemical analysis of FRa expression
Immunohistochemical (IHC) analysis was performed in the human pituitary
tumors to determine the cellular and subcellular localization of FRa
overexpression. Frozen tissue sections from five NF, six NF, two PRL,
three GH, and two ACTH-secreting adenomas and two normal anterior
pituitary glands were used for these analyses. Three specimens of ovarian
Folic acid binding in pituitary adenomas
*
**
***
Tumor type
n = 5
Control
0
50
F
o
l
i
c
a
c
i
d
b
i
n
d
i
n
g
(
p
m
o
l
/
m
g
)
100
150
10
NF
13
NF+
5
PRL
5
GH
4
ACTH
Figure 8.4 Box plots representing folic acid binding to FRa by adenomas subtypes.
A horizontal line in each box represents the median value of folic acid binding to FRa of
each group. Box, the 25th and 75th percentile range of scores. Whiskers, the highest and
lowest values. Numbers of pituitaries tested in each group were n 5 for normal
pituitary controls, 10 for NF, 13 for NF, 5 for PRL, 5 for GH, and 4 for ACTH-
secreting adenomas. *, in NF adenomas, folate binding was significantly different
from controls ( p 0.007), PRL ( p 0.003), GH ( p 0.003), and ACTH-secreting
adenomas ( p 0.002). **, in NF adenomas, folate binding was significantly different
from controls ( p 0.001), PRL ( p 0.001), GH ( p 0.001), and ACTH-secreting
adenomas ( p 0.001). ***, in PRL adenomas, folate binding was significantly different
from controls ( p 0.047).
Folate Receptor Expression in Pituitary Adenomas 247
Figure 8.5 IHC demonstrating cytoplasmic FRa in anterior pituitary gland and
pituitary adenomas. Ovarian adenocarcinoma served as a positive control for FRa
receptor IHC. Strong immunoreactivity for FRa was present on the luminal membrane
of ovarian adenocarcinoma glands and, to a lesser extent, within the cytoplasm(A; arrow).
Ovarian adenocarcinoma did not show any immunoreactivity when the primary antibody
directed against FRa was replaced with PBS (B; negative control). Normal (nonneoplas-
tic) anterior pituitary gland showed focal weak staining for FRa in glandular cells but
248 Chheng-Orn Evans et al.
adenocarcinoma served as a positive control for FRa IHC because this
tumor type consistently expresses high levels of FRa. All of the ovarian
adenocarcinomas showed strong luminal and membranous immunoreactiv-
ity for FRa and moderate cytoplasmic staining (Fig. 8.5A, arrow). The
ovarian adenocarcinomas showed no immunoreactive staining when the
primary antibody (LK26) FRa was replaced with normal PBS (Fig. 8.5B),
indicating that there was no nonspecific staining from the secondary anti-
body. All five of the NF adenomas showed FRa expression by IHC.
Immunostaining was seen diffusely in the cytoplasm of tumor cells in these
adenomas and varied from strong in four tumors (100% of cells; Fig. 8.5D)
to moderate in one tumor (50% of cells; Fig. 8.5E). No staining was noted
within other cellular constituents of the adenoma such as vascular structures
or supporting stromal elements. The six NF adenomas that showed focal
weak staining for hormones (Fig. 8.5F, stained for LH, arrow) expressed high
levels of FRa (Fig. 8.5G, stained with LK26). Strong cytoplasmic expression
was also seen in five of these six NFtumors (100% of cells), and more than
moderate staining was seen in one tumor (80% of cells). Cytoplasmic FRa
expression in these NF adenomas was much greater than that seen in normal
(nonneoplastic) anterior pituitary glands, which showed only focal weak
staining of pituitary cells (Fig. 8.5C, arrow).
In the seven functional adenomas analyzed (two PRL, two ACTH, and
three GH adenomas), FRa expression was either totally absent (two GH
tumors, 0% of cells) or only minimally detected (15% of cells in two PRL
and three other tumors). Pituitary adenomas that showed strong staining for
PRL hormone (Fig. 8.5H) showed very light staining for FRa (Fig. 8.5I,
stained with LK26).
In this study, we have following results.
F. Selection of clones of aT3-1 cells stably expressing FRa
and mFRa
Cells (aT3-1) were transfected with expression vector containing FRa
cDNA (pZeoSV-FRa) or mutant FRa 67 (pZeoSV-FR67) or expression
vector alone (pZeoSV). After 3 weeks in 300 mg/ml Zeocin, 30 clones from
each of pZeoSV-FRa, pZeoSV-FR67, and pZeoSV-transfected cells were
none in stromal cells (C; arrow). NF adenomas that did not show any evidence of
hormone production by IHC were all immunoreactive for FRa, with the intensity of
staining various from strong (D) to moderate (E). Immunoreactivity for FRa was limited
to the cytoplasm of adenoma cells and not present in stromal cells or blood vessels. NF
adenomas that showed only focal, weak immunoreactivity for hormones (IHC for LH
shown by arrow in panel F) also showed strong cytoplasmic expression FRa (G). Pituitary
adenomas such as a prolactinoma showing cytoplasmic immunoreactivity for PRL by IHC
(H), but did not show appreciable cytoplasmic FRa (I).
Folate Receptor Expression in Pituitary Adenomas 249
selected and expanded. Three representative clones for pZeoSV-FRa
(named as FRwt-3, FRwt-11, and FRwt-14) and mutant FRa 67 (named
as FR67-9, FR67-11, and FR67-27), one clone from pZeoSV (pZeo) were
selected for further studies. Selection of clones was based on the protein
expression of FRa protein. Figure 8.6 demonstrated the protein expression
of these clones. Similar endogenous FRa protein expression was exhibited
in pZeo cells as that of aT3-1 cells. The clones transfected with pZeoSV-
FRa (FRwt-3, FRwt-11, and FRwt-14) showed high level of FRa
protein expression while the mutant FR67 clones (FR67-9, FR67-11,
and FR67-27) showed moderate high level of FRa protein expression
compared to the endogenous expression of parent aT3-1 cells and pZeo
cells. b-Actin protein revealed similar expression level in all the cells tested.
G. FRa induces cell proliferation in aT3-1 cells
To determine whether FRa affects cell proliferation, all clones were plated
at 1 10
4
/ml in 24-well plate in either folic acid-free medium or regular
medium with folic acid and were cultured for 10 days. When cultured in
regular medium containing 10 mM of folic acid, the clone transfected with
either FRa or FR67 mutant presented no difference in growth rate com-
paring with cells transfected with vector alone or parent aT3-1 cells (data
not shown). While cultured in folic acid-free medium containing 10 nM of
folic acid resembling physiological concentration of folic acid, three clones
of FRa-transfected cells (FRwt-3, FRwt-11, and FRwt-14) displayed
significantly greater proliferation rate than the cells transfected with control
vector only at all time points tested (Fig. 8.7). The three transfected clones
with FR67 mutant exhibited slightly reduced proliferation rate than the
cells transfected with vector only and significantly reduced growth rate
compared with FRwt clones (Fig. 8.7). These experiments indicate that
overexpression of FRa does induce a significant proliferation effect in aT3-
1 cell, while overexpression of mutant FR67 counteracts the proliferation
effect of FRa in these cells.
a
T
3
-
1
p
Z
e
o
-
5
F
R
w
t
-
3
F
R
w
t
-
1
1
F
R
w
t
-
1
4
F
R
6
7
-
9
F
R
6
7
-
1
1
F
R
6
7
-
2
7
FRa
b-Actin
Figure 8.6 Representative Western blot analysis of FRa protein (A) and b-actin
protein (B) expression in aT3-1 cells, pZeo-transfected cells, FRa-transfected clones
(FRwt-3, FRwt-11, FRwt-14) and mutation in FR67-transfected clones (FR67-9,
FR67-11, FR67-27).
250 Chheng-Orn Evans et al.
H. FRa induces cell cycle progression and PCNA expression
The cell cycle profile was determined by a BrdU incorporation assay to
address the effect of overexpression of FRa in the context of cell cycle
regulation. Cells (aT3-1) overexpressing FRa presented an increase in cells
residing in S-phase (1822%) compared with 1314% in cells transfected
with pZeo and parental cells ( p < 0.05). On the other hand, the cells
transfected with mutant FR67 displayed a decreased number of cells resid-
ing in S-phase (1012%), while no statistical significance was observed
compared with control cells. These results indicate that FRa induces cell
cycle progression in S-phase in aT3-1 cells (Table 8.4). To further confirm
that FRa overexpression influences cell proliferation, PCNA staining using
flow cytometry technique was measured in aT3-1 and derived clones.
PCNA is a nuclear antigen that is only expressed in proliferating cells and
absent in resting cells. Cells transfected with FRa (FRwt-3, FRwt-11, and
FRwt-14) exhibited higher level of PCNA expression compared with pZeo
clones and parental aT3-1 cells. On the contrary, cells transfected with
mutant FR67 showed similar or slightly lower lever of PCNA expression
compared with control cells (Fig. 8.8).
I. FRa promotes growth in soft agar assay
Cells (aT3-1) and all clones derived from aT3-1 were further tested for the
anchorage-independent growth in soft agar assay. This assay was performed
by seeding 5000 cells in a 35 mm culture dish and the number of colonies
0
10
20
30
40
0 2 4 6 8 10
Days
C
e
l
l
n
u
m
b
e
r
s
(
1
0
4
)
FRwt-11
FRwt-14
FRwt-3
pZeo
Cell
FR67-27
FR67-11
FR67-9
Figure 8.7 Overexpression of FRa induces cell growth in aT3-1 cells. Cells (aT3-1)
and single clone of stable cells overexpressing FRa and mutant FR67, at an initial
concentration of 0.2 10
5
/ml, were harvested at time points indicated and counted.
Shown are mean values SD from three independent experiments, each performed in
triplicate.
Folate Receptor Expression in Pituitary Adenomas 251
larger than 20 cells formed in 2 weeks was counted. The cell growth in soft
agar was significantly increased in FRa overexpressing cells compared with
parent and pZeo cells ( p < 0.001). The cells transfected with mutant FR67
formed a significantly lower number of colonies compared with parent and
pZeo cells ( p <0.01) (Fig. 8.9). These results suggest that overexpression of
FRa in aT3-1 cells may induce cellular transformation, and the FR67
mutation abrogates this ability of FRa.
J. [
3
H]Folic acid binding
In order to determine the total amount of FRa in each cell type, the
solubilized folic acid-binding assay of each cell type was performed in two
independent experiments, each sample was done in duplicate. JAR cells and
NIH3T3 cells were used, respectively, as positive and negative control cells.
Total solubilized folic acid binding indicated that the three FRwt cells had
significantly higher binding of folic acid than the FR67 mutant, the pZeo,
and the aT3-1 cell (Fig. 8.10).
K. FRa induces aT3-1 cell growth through NOTCH3 pathway
RT-qPCR was performed to investigate the possible mechanisms involved
in induced cell proliferation by FRa overexpression in aT3-1 cells. We are
interested in studying whether FRa associates with the recognized pathways
in pituitary adenoma tumorigenesis. PTTG1, FGFR1, and estrogen recep-
tors have been proposed important roles in pathogenesis of pituitary ade-
noma. Elevated PTTG1 and FGFR1, EGFR (epidermal growth factor
receptor), and downregulated estrogen receptor have been discovered in
NF pituitary tumors (Bradshaw and Kakar, 2007; Chaidarun et al., 1994;
Gittoes, 1998; McCabe et al., 2002, 2003; Onguru et al., 2004). In our
Table 8.4 BrdU incorporation assay in folate receptor overexpressing cells
G0/G1-phase (%) S-phase (%) G2-phase (%)
aT3 78 3 14 2 8 2
pZeo 77 3 13 2 10 1
FRwt-3 71 5 20 2* 10 3
FRwt-11 67 4 22 3* 11 1
FRwt-14 70 2 18 3* 12 2
FR67-9 83 3 10 1 7 1
FR67-11 80 5 12 3 8 2
FR67-27 81 2 11 2 8 2
* p < 0.05, compared with untransfected aT3-1 cells and pZeo cells.
Mean values are from two independent experiments, each done in triplicate.
252 Chheng-Orn Evans et al.
previous report, we proposed that activated Wnt pathway and NOTCH3
pathway may be engaged in NF pituitary adenoma pathogenesis (Moreno
et al., 2005). Based on these observations, the genes selected for this study
are genes involved in Wnt pathway including SFRP1, b-catenin, PITX2,
cyclin D1, RB1, and TLE2, and the genes related to NOTCH3 pathway
including NOTCH3, ASCL-1, DLK1, HES-1, and PIT1. Other genes
involved in the pathogenesis of NF adenoma including EGFR, FGFR1
(fibroblast growth factor receptor 1), ERa (estrogen receptor a), ERb
(estrogen receptor b), PTTG1 were also examined.
FRa mRNA overexpression was confirmed by RT-qPCR in both
FRwt clones and FR67 clones, the two groups of cells displayed similar
mRNA FRa expression (data not shown). Parent aT3-1 and pZeo cells
49.00 0.06
58.80
48.57
38.75
59.92 62.00
46.70 44.65
H
i
s
t
o
g
r
a
m
FRwt-3
FL1-H: PCNA-FITC
10
0
50
100
150
200
250
10
1
10
2
10
3
10
4
H
i
s
t
o
g
r
a
m
FRwt-11
FL1-H: PCNA-FITC
10
0
50
100
150
200
250
10
1
10
2
10
3
10
4
H
i
s
t
o
g
r
a
m
FRwt-14
FL1-H: PCNA-FITC
10
0
50
100
150
200
250
10
1
10
2
10
3
10
4
H
i
s
t
o
g
r
a
m
Negative control
FL1-H: PCNA-FITC
10
0
10
20
30
40
10
1
10
2
10
3
10
4
H
i
s
t
o
g
r
a
m
T3-1
FL1-H: PCNA-FITC
10
0
100
200
300
10
1
10
2
10
3
10
4
H
i
s
t
o
g
r
a
m
pZeo
FL1-H: PCNA-FITC
10
0
100
200
300
10
1
10
2
10
3
10
4
H
i
s
t
o
g
r
a
m
FR67-9
FL1-H: PCNA-FITC
10
0
50
100
150
200
250
10
1
10
2
10
3
10
4
H
i
s
t
o
g
r
a
m
FR67-11
FL1-H: PCNA-FITC
10
0
100
200
300
10
1
10
2
10
3
10
4
H
i
s
t
o
g
r
a
m
FR67-27
FL1-H: PCNA-FITC
10
0
50
100
150
200
250
10
1
10
2
10
3
10
4
Figure 8.8 Flow cytometric measurement of PCNA in stably transfected and nontrans-
fected aT3-1 cells. (A) The negative control was stained with the FITC-conjugated
isotype control mAb; (B) Nontransfected aT3-1 cells; (C) pZeo cells; (D) FRwt-3
cells; (E) FRwt-11 cells; (F) FRwt-14 cells; (G) FR67-9 cells; (H) FR67-11 cells;
(I) FR67-27 cells. The results shown are representative of three different experiments.
Folate Receptor Expression in Pituitary Adenomas 253
T3-1
FRwt-3 FRwt-11 FRwt-14
pZeo
FR67-9 FR67-27 FR67-11
B
0
50
100
150
200
250
T3 pZeo FRwt-3 FRwt-11 FRwt-14 FR67-9 FR67-11 FR67-27
C
o
l
o
n
y
n
u
m
b
e
r
s
A
p < 0.01
p < 0.001
Figure 8.9 Overexpression of FRa in aT3-1 cells induces soft agar formation. Control
clone (aT3-1, pZeo), FRa overexpressing clones, and mutant FR67 overexpressing
clones were seeded in triplicate into 0.3% agar in six-well plates at 5 10
3
/ml.
(A) After 15 days of culture, colonies containing more than 20 cells were enumerated.
Results are mean SD for three separate experiments. (B) After 15 days of culture, the
plates were counted and photographed. Representative results from three independent
experiments are shown.
254 Chheng-Orn Evans et al.
showed no ectopic FRa mRNA expression (data not shown). Three groups
of cells (FRwt, FR67, and control cells) demonstrated similar level of
endogenous FRa expression by using murine cDNA primers (data not
shown).
Interestingly, NOTCH3 mRNA expression was significantly induced in
FRwt clones ( p < 0.001) compared to that of pZeo clone. The expression
of NOTCH3 mRNA was decreased in mutant FR67 clones, but no
statistical difference was observed between FR67 group and pZeo clone
(Fig. 8.11A). HES-1 mRNA was increased in FR67 mutant clones ( p <
0.01, comparing to pZeo cells) (Fig. 8.11B) but not in FRwt clones. No
detectable mRNAs expression of DLK1 and PIT1 were observed in these
three groups of cells. There was no marked difference of the ASCL-1
expression among these cells (data not shown).
FGFR1 mRNA expression was increased in FRwt clones ( p < 0.05)
compared to that in pZeo. No significant change of FGFR1 expression was
found in FR67 mutant clones (Fig. 8.11C). While ESR1 mRNA expres-
sion showed no difference among pZeo, FRwt, and FR67 cells, the
substantially reduced ESR2 mRNA expression was observed in FRwt
clones ( p < 0.001). A lesser degree of reduced ESR2 mRNA expression
was detected in FR67 mutant clones ( p < 0.01) compared to that of pZeo
(Fig. 8.11D). However, no substantial changes of PTTG1 and EGFR were
observed in either FRwt clones or FR67 mutant clones compared to the
control cells (data not shown).
However, most of the genes involved in Wnt pathway, including
SFRP1, b-catenin, PITX2, cyclin D1, and RB1 mRNA expression
demonstrated no significant difference among three groups of cells. While
TLE2 mRNA expression exhibited an increasing tendency in all FRwt
clones compared to pZeo, no statistical difference was noticed (Fig. 8.11E).
Surprisingly, a substantial higher expression of TLE mRNA was observed in
Folic acid binding
5
0
5
10
15
20
25
30
aT3-1 pZEO FRwt-3 FRwt-11 FRwt-14 FR67-9 FR67-11 FR67-27
S
p
e
c
i
f
i
c
b
i
n
d
i
n
g
(
p
m
o
l
/
m
g
)
*
*
*
Figure 8.10 Solubilizedfolic acidbinding. Solubilizedfolic acidbinding was as described
inMaterials andMethods (SectionH). Meanvalueare fromtwoindependent experiments.
*p <0.05, compare with untransfected aT3-1, transfected pZeo, and FR67 cells.
Folate Receptor Expression in Pituitary Adenomas 255
FR67 mutant clones compared to that in pZeo cells ( p < 0.05). TLE2
mRNA resembled HES-1 mRNA in their expression pattern among all the
cells studied.
0
1
2
3
4
5
6
7
pZeo FRwt-3 FRwt-11 FRwt-14 FR67-9 FR67-11 FR67-27
N
o
t
c
h
3
m
R
N
A
e
x
p
r
e
s
s
i
o
n
(
f
o
l
d
c
h
a
n
g
e
)
A
Tubulin
Actin
p<0.001
ND
pZeo FRwt-3 FRwt-11 FRwt-14 FR67-9 FR67-11 FR67-27
B
0
1
2
3
4
5
H
E
S
-
1
m
R
N
A
e
x
p
r
e
s
s
i
o
n
(
f
o
l
d
c
h
a
n
g
e
)
Tubulin
Actin
P<0.01
ND
Figure 8.11 (continued)
256 Chheng-Orn Evans et al.
0.0
0.5
1.0
1.5
2.0
0.00
0.50
1.00
1.50
2.00
D
C
F
G
F
R
1
m
R
N
A
e
x
p
r
e
s
s
i
o
n
(
f
o
l
d
c
h
a
n
g
e
)
Tubulin
Actin
Tubulin
Actin
P<0.05
ND
E
S
R
2
m
R
N
A
e
x
p
r
e
s
s
i
o
n
(
f
o
l
d
c
h
a
n
g
e
)
p<0.05
pZeo FRwt-3 FRwt-11 FRwt-14 FR67-9 FR67-11 FR67-27
pZeo FRwt-3 FRwt-11 FRwt-14 FR67-9 FR67-11 FR67-27
(continued)
Folate Receptor Expression in Pituitary Adenomas 257
L. In vivo imaging of FR expression
SPECT/CT of FolateScan (Technetium Tc-99m EC20) in pituitary tumors
is a novel imaging tracer and technique for in vivo imaging of FR expression
in clinically NF pituitary tumors that express the FR. The folate-targeted
imaging agent FolateScan (Technetium Tc-99m EC20; Endocyte, Inc.)
consists of technetium-99m conjugated to an FR ligand (EC20) in clinically
NF pituitary tumors. Because the FR is overexpressed in clinically NF
tumors, uptake of FolateScan in the tumor can serve as a marker for selection
of FR tumors for folate-targeted therapy.
To perform the study, each patient had a medical examination, met
eligibility criteria, and signed an informed consent in an IRB-approved
protocol. Baseline vital signs, blood, and urine samples were obtained for
analysis 2 h before injection. Patients received two IV injections 13 min
apart: 1 mg of folic acid and 12 ml injection of 0.1 mg/666 MBq of
FolateScan. The patients had midthigh to head planar images at 12 h
postinjection followed by imaging using a SPECT scanner with integrated
CT (GE VG/Hawkeye). Following surgery, Western blot analysis was
performed to detect folate-receptor expression in tumor samples.
Patients were imaged prior to surgery, and their planar and SPECT
images were compared to Western blot analysis. Patients had presented with
visual deficit and a hormonal profile consistent with clinically NF pituitary
0.00
0.50
1.00
1.50
2.00
2.50
3.00
3.50
T
L
E
2
m
R
N
A
e
x
p
r
e
s
s
i
o
n
(
f
o
l
d
c
h
a
n
g
e
)
Tubulin
Actin
*
P < 0.05 (TLE2/tubulin); p < 0.01(TLE2/actin) E
ND
*
Wt-11 vs pZeo p < 0.05 (TLE2/actin)
pZeo FRwt-3 FRwt-11 FRwt-14 FR67-9 FR67-11FR67-27
Figure 8.11 mRNAs expression of NOTCH3 (A), HES-1 (B), FGFR1 (C), ESR2 (D),
and TLE2 (E) in FRa overexpressing and mutant FR67 overexpressing cells compared
with control pZeo cells. All measurements are shown relative to the expression levels of
actin and tubulin. Values are based on three independent experiments, each performed
in duplicate.
258 Chheng-Orn Evans et al.
tumor. FolateScan successfully indentifies the FR pituitary tumor. The
Western blot analysis validates that this tumor was FR positive. Similarly
in patients where the tumor was FR by FolateScan scintigraphy, this was
validated by lack of FR expression by Western blot analysis. These prelimi-
nary data demonstrate that FolateScan can successfully target FR pituitary
tumors and provide preliminary evidence that folate can be used as a delivery
system for tumor-targeted delivery of drugs into pituitary tumors (Figs. 8.12
and 8.13).
IV. Discussion
Our result demonstrates that FRa overexpression induces growth in
the NF pituitary tumor cell line aT3-1 in physiological folic acid culture
condition and soft agar. The increased cell growth in FRwt cells were
further confirmed by BrdU incorporation assay (BrdU is only incorporated
into the DNA of proliferating cells) and PCNA staining (PCNA is a nuclear
antigen that is only expressed in proliferating cells and absent in resting
cells). These results indicate that FRa overexpression may confer a growth
advantage to pituitary tumor cells. Our data are in agreement with the
observation that transfection of FRa in NIH/3T3 cells is associated with an
Tc-99m FolateScan
Protocol
2 IV injections 13 minutes
apart:
(1) 1 mg folic acid
(2) 12 mL/0.1 mg
Folatescan
(EC20) labeled with 1525
mCi Tc-99m
12 h postinjection:
Whole body imaging
SPECT/CT images follow
whole body
Tc-99m Folatescan whole body
Patient 1: Folate-Receptor positive
Patient 2: Folate-Receptor negative
Gadolinium-enhanced MRI
Tc-99m Folatescan whole body
Gadolinium-enhanced MRI
Figure 8.12 SPECT/CT of Tc-99m FolateScan in pituitary tumors. A novel imaging
tracer and technique.
Folate Receptor Expression in Pituitary Adenomas 259
increased growth rate both in vitro and in vivo (Bottero et al., 1993). Our data
are further supported by the study in the epithelial ovarian cancer that
demonstrates that FRa overexpression is significantly correlated with the
percentage of cells in the S-phase (Toffoli et al., 1997). Therefore, FRa
overexpression may present a stimulus for cell growth in NF pituitary
tumors.
The degree of FRa expression is associated with biological aggressive-
ness of ovarian neoplasms and suggests an involvement of FRa in neoplastic
progression (Toffoli et al., 1997). Functional downregulation of FRa with
specific intracellular expression of single-chain antibodies (intrabody) in
ovarian cancer cells was accompanied by reduced cell proliferation and
adhesion also supports the role of FRa in cancer progression (Figini et al.,
2003). Antisense oligonucleotides targeted to the FRa induced a dose-
dependent decrease in breast cancer cell survival. Jhaveri et al. (2004)
indirectly confirms the importance of FRa in cancer development. Alternate
mechanisms of FRa enhancement of cell proliferation have been eluci-
dated; for instance, FRa is partitioned in cellular-membrane domains in
close physical and functional association with the src-family member p53
56 lyn and the Gai-3 subunit of heterotrimeric G-proteins (Miotti et al.,
2000) and FRa expression transcriptionally regulates the expression of the
tumor suppressor gene caveolin-1 (Bagnoli et al., 2000). Upregulated FRa
has been associated with two pathways by statistical analysis of RT-qPCR
Folate-receptor positive
10/1 SPECT in target/background
CT correlation confirms uptake in pituitary
Folate-receptor negative
1/1 SPECT in target/background
CT correlation guides search for uptake
Western blot analysis FR Western blot analysis FR+
10 mg 20 mg 10 mg 20 mg
Coronal
Sagittal
Coronal
Sagittal
Patient 1 Patient 2
Figure 8.13 SPECT/CT of Tc-99m FolateScan in pituitary tumors. A novel imaging
tracer and technique. FolateScan can be used to image FR expression in nonfunctional
FR pituitary tumors and has potential as a mechanism for tumor-targeted drug/
therapy delivery.
260 Chheng-Orn Evans et al.
data in a panel of 39 microdissected ovarian carcinoma (Hough et al., 2001).
However, how FRa interacts with the genes in those two pathways has not
been examined. In another experiment, the opposite effect has been
observed in the transduction of cervical carcinoma cells with FRa cDNA,
where FRa overexpression inhibits cell proliferation (Sun et al., 1995,
1999). In a previous study, we have demonstrated that FRa is overexpressed
in NF pituitary tumors (Evans et al., 2001, 2003; Moreno et al., 2005);
however, whether the perturbation of FRa gene expression is involved
pituitary tumor cell growth is unknown.
The mutant FR67 was used in this study because it was discovered to
inhibit folate binding and uptake in MA104 cells, a monkey kidney epithe-
lial cell line, and confer a dominant negative phenotype (Orr and Kamen,
1994, 1995). Our study showed that FRa protein expression was decreased
in mutant FR67 (see Fig. 8.6). Additionally, we discovered that mutant
FR67 slowed cell growth in cell culture and remarkably reduced colony
formation in soft agar as compared with wild-type FR. Similarly, mutant
FR67 showed remarkably decreased numbers of cells residing in S-phase by
BrdU assay and reduced PCNA staining as compared with FRwt, and
slightly delayed S-phase progression and decreased PCNA staining as com-
pared with control cells.
It is suggested that elevated levels of FRa induce cell proliferation not
only by mediating folate uptake, which in turn, facilitates DNA synthesis
and replication and induces cell proliferation, but also by generating other
regulating signals. Our study found that FRa overexpressing cells had folic
acid-binding capacity significantly higher than both control cells (aT3-1
and pZeo). However, the folic acid-binding capacity in FR mutant cells was
significantly decreased ( p 0.001) when compared to FRa overexpressing
cells. Perhaps this may be part of the mechanism that FRa induces cell
proliferation in aT3-1 cells.
Tumorigenicity was determined by transplanting tumor cells subcuta-
neously in the flank of nude mice. No obvious tumor growth was observed
in any of the mice when each group of four immunodeficient mice was in-
jected with (5 10
6
cells) either FRa (FRwt-3, FRwt-11, and FRwt-14),
mutant FR67 (FR67-9, FR67-11, and FR67-27), and pZeo-transfected
and parental aT3-1 cells.
It has also been proposed that FRa may promote cell proliferation
independent of folate internalization. For example, hFR, similar to other
glycosylphosphatidylinositol-linked proteins (Horejsi et al., 1999;
Ilangumaran et al., 2000), may have a role in cell signaling processes leading
to cell growth and proliferation (Miotti et al., 2000). Our data suggest that
FRa may be associated with tissue-specific regulatory signals such as
NOTCH3 and FGFR1 to regulate cell growth.
Emerging data support the role of the NOTCH signaling pathway
in tumorigenesis and neural development. Constitutive expression of
Folate Receptor Expression in Pituitary Adenomas 261
NOTCH3-IC in the peripheral epithelium in the developing lung of
transgenic mice resulted in altered lung morphology and delayed develop-
ment, suggesting that NOTCH3 signaling could contribute to lung cancer
progression through the inhibition of terminal differentiation (Dang et al.,
2003). Moreover, transgenic mice expressing NOTCH3-IC in thymocytes
and T cells developed early and aggressive T-cell neoplasias (Bellavia et al.,
2002). NOTCH3 has been extensively studied in brain development, and
recently it has been functioned as an oncogene in the brain (Dang et al.,
2006). In our study, by RT-qPCR, we discovered that NOTCH3 mRNA
is significantly upregulated in FRwt cells and slightly reduced in FR67 cells
as compared with control pZeo cells. This observation, in conjunction with
our previous report that demonstrated that both NOTCH3 and FRa were
overexpressed in NF pituitary adenoma (Moreno et al., 2005), indicates that
FRa may correlate with NOTCH3, a brain oncogene, in promoting cell
growth and tumor formation in pituitary tumors.
A high level of mRNA and protein expression of FGFR1 has been
observed in pituitary adenomas compared with normal pituitaries. Signifi-
cantly enhanced expression of FGFR1 was observed in invasive adenomas
compared with other pituitary tumors, suggesting the receptor-mediated
mechanisms of growth factor action may be critically important in pituitary
tumorigenesis (McCabe et al., 2003). Our previous microarray study in 23
NF pituitary tumors also found increased FGFR1 expression (Moreno et al.,
2005). However, with immunohistochemistry, some researchers found that
only basic fibroblast growth factor was increased in pituitary adenoma and
the FGFR1 immunoreactivity was inversely correlated with maximum
tumor diameter (Fukui et al., 2002). By RT-qPCR, we found that
FGFR1 mRNA was upregulated in FRwt-transfected cells and ESR2
mRNA was downregulated in FRwt cells.
ERb mRNA expression was detected in the adult normal human pitui-
tary gland; however, the expression of ERb in NF pituitary tumors was
reduced when compared with normal pituitaries (Gittoes et al., 1999). This
is consistent with our microarray finding that showed decreased ERb
expression in NF human pituitary tumors. In breast cancer, a negative
correlation has been observed between ER and FRa (Rochman et al.,
1985). More recently, it has been demonstrated that ER represses both
the FRa promoter and the endogenous FRa gene expression (Kelley et al.,
2003). Our data shows that overexpression of FRa downregulates ESR2
may indicate that there may be a two-way communication between FRa
and ESR2 in pituitary tumor cells. However, whether FRa directly
or indirectly interacts with NOTCH3, FGFR1, and ESR2 requires further
investigation.
Abnormal nuclear accumulation and missense mutations of b-catenin
have been observed in 57% pituitary tumor patients, suggesting that the
upregulation of the Wnt signaling pathway plays an important role in the
262 Chheng-Orn Evans et al.
tumorigenesis and development of pituitary adenomas (Semba et al., 2001).
In our previous study, we observed mRNA overexpression in PITX2,
cyclin D1, and NF pituitary adenomas. This indicates that an elevated
Wnt/b-catenin signaling may be involved in the progression of these
tumors (Moreno et al., 2005). However, RT-qPCR of most of these
genes related to Wnt pathway fails to show mRNA alteration of SFRP1,
b-catenin, and cyclin D1. TLE2 (transducin-like enhancer of split 2), a
mammalian homologue of the Drosophila transcriptional repressor Groucho,
displayed increased expression in FR67 clones and showed a similar expres-
sion pattern as HES-1 in pZeo, FRwt, and FR67 cells. TLE2 was found to
interact and is coexpressed with HES-1 in both neural and nonneural tissues
(Grbavec et al., 1998). Furthermore, the universal Groucho/TLE corepres-
sors have also been implicated as components of the NOTCH pathway and
effectors of various other signaling cascades including the nuclear transdu-
cers of Wingless/Wnt signaling pathways, EGFR signaling, and other
receptor tyrosine kinase pathways (Hasson and Paroush, 2006). In our
previous communication, we proposed that TLE2 may serve as a link
between NOTCH3 pathway and Wnt pathway. While in this report,
TLE2 was found to be the only gene affected by FRa overexpression in
Wnt pathway. It is possible that TLE2 may act as a common cellular target
for cross-pathway regulation by FRa between NOTCH3 pathway and
other pathways.
In conclusion, our data suggest that FRa overexpression facilitates cell
proliferation and plays an important role in pituitary tumorigenesis. Poten-
tially, this finding could be exploited in order to develop new, innovative
molecular targeted treatment modalities for NF pituitary adenomas.
ACKNOWLEDGMENTS
We gratefully acknowledge financial assistance to Nelson M. Oyesiku, MD, PhD, FACS,
from the National Institutes of Health (R01-NS5143901). We thank the Department of
Neuropathology, Emory University Hospital, for the histology and IHC analysis.
REFERENCES
Antony, A. C. (1992). The biological chemistry of folate receptors. Blood 79, 28072820.
Asa, S. (1998). Tumors of the Pituitary Gland. Armed Forces institute of Pathology,
Washington, DC.
Asa, S. L., and Ezzat, S. (1998). The cytogenesis and pathogenesis of pituitary adenomas.
Endocr. Rev. 19, 798827.
Asa, S. L., and Kovacs, K. (1992). Clinically non-functioning human pituitary adenomas.
Can. J. Neurol. Sci. 19, 228235.
Asa, S. L., Cheng, Z., Ramyar, L., Singer, W., Kovacs, K., Smyth, H. S., and Muller, P.
(1992). Human pituitary null cell adenomas and oncocytomas in vitro: Effects of
Folate Receptor Expression in Pituitary Adenomas 263
adenohypophysiotropic hormones and gonadal steroids on hormone secretion and tumor
cell morphology. J. Clin. Endocrinol. Metab. 74, 11281134.
Bagnoli, M., Tomassetti, A., Figini, M., Flati, S., Dolo, V., Canevari, S., and Miotti, S.
(2000). Downmodulation of caveolin-1 expression in human ovarian carcinoma is
directly related to alpha-folate receptor over-expression. Oncogene 19, 47544763.
Bellavia, D., Campese, A. F., Checquolo, S., Balestri, A., Biondi, A., Cazzaniga, G.,
Lendahl, U., Fehling, H. J., Hayday, A. C., Frati, L., von Boehmer, H., Gulino, A.,
et al. (2002). Combined expression of pTalpha and Notch3 in T cell leukemia identifies the
requirement of preTCR for leukemogenesis. Proc. Natl. Acad. Sci. USA 99, 37883793.
Black, P. M., Hsu, D. W., Klibanski, A., Kliman, B., Jameson, J. L., Ridgway, E. C.,
Hedley-Whyte, E. T., and Zervas, N. T. (1987). Hormone production in clinically
nonfunctioning pituitary adenomas. J. Neurosurg. 66, 244250.
Bottero, F., Tomassetti, A., Canevari, S., Miotti, S., Menard, S., and Colnaghi, M. I. (1993).
Gene transfection and expression of the ovarian carcinoma marker folate binding protein
on NIH/3T3 cells increases cell growth in vitro and in vivo. Cancer Res. 53, 57915796.
Bradshaw, C., and Kakar, S. (2007). Pituitary tumor transforming gene: An important gene
in normal cellular functions and tumorigenesis. Histol. Histopathol. 22, 219226.
Chaidarun, S. S., Eggo, M. C., Sheppard, M. C., and Stewart, P. M. (1994). Expression of
epidermal growth factor (EGF), its receptor, and related oncoprotein (erbB-2) in human
pituitary tumors and response to EGF in vitro. Endocrinology 135, 20122021.
Chung, K. N., Saikawa, Y., Paik, T. H., Dixon, K. H., Mulligan, T., Cowan, K. H., and
Elwood, P. C. (1993). Stable transfectants of human MCF-7 breast cancer cells with
increased levels of the human folate receptor exhibit an increased sensitivity to antifolates.
J. Clin. Invest. 91, 12891294.
Dang, L., Fan, X., Chaudhry, A., Wang, M., Gaiano, N., and Eberhart, C. (2006). Notch3
signaling initiates choroid plexus tumor formation. Oncogene 25, 487491.
Dang, T. P., Eichenberger, S., Gonzalez, A., Olson, S., and Carbone, D. P. (2003).
Constitutive activation of Notch3 inhibits terminal epithelial differentiation in lungs of
transgenic mice. Oncogene 22, 19881997.
Doucette, M. M., and Stevens, V. L. (2001). Folate receptor function is regulated in response
to different cellular growth rates in cultured mammalian cells. J. Nutr. 131, 28192825.
Evans, C. O., Young, A. N., Brown, M. R., Brat, D. J., Parks, J. S., Neish, A. S., and
Oyesiku, N. M. (2001). Novel patterns of gene expression in pituitary adenomas
identified by complementary deoxyribonucleic acid microarrays and quantitative reverse
transcription-polymerase chain reaction. J. Clin. Endocrinol. Metab. 86, 30973107.
Evans, C. O., Reddy, P., Brat, D. J., ONeill, E. B., Craige, B., Stevens, V. L., and
Oyesiku, N. M. (2003). Differential expression of folate receptor in pituitary adenomas.
Cancer Res. 63, 42184224.
Figini, M., Ferri, R., Mezzanzanica, D., Bagnoli, M., Luison, E., Miotti, S., and Canevari, S.
(2003). Reversion of transformed phenotype in ovarian cancer cells by intracellular
expression of anti folate receptor antibodies. Gene Ther. 10, 10181025.
Fukui, S., Otani, N., Nawashiro, H., Yano, A., Nomura, N., Miyazawa, T., Ohnuki, A.,
Tsuzuki, N., Katoh, H., Ishihara, S., and Shima, K. (2002). Subcellular localization of
basic fibroblast growth factor and fibroblast growth factor receptor 1 in pituitary adeno-
mas. Brain Tumor Pathol. 19, 2329.
Gao, Y., Qian, W., Dark, K., Toraldo, G., Lin, A., Guldberg, R., Flavell, R.,
Weitzmann, M., and Pacifici, R. (2004). Estrogen prevents bone loss through transform-
ing growth factor beta signaling in T cells. Proc. Natl. Acad. Sci. USA 101, 1661816623.
Gittoes, N. J. (1998). Current perspectives on the pathogenesis of clinically non-functioning
pituitary tumours. J. Endocrinol. 157, 177186.
Gittoes, N. J., McCabe, C. J., Sheppard, M. C., and Franklyn, J. A. (1999). Estrogen receptor
beta mRNA expression in normal and adenomatous pituitaries. Pituitary 1, 99104.
264 Chheng-Orn Evans et al.
Grbavec, D., Lo, R., Liu, Y., and Stifani, S. (1998). Transducin-like enhancer of split 2,
a mammalian homologue of Drosophila Groucho, act as a transcriptional repressor, interacts
with hairy/enhancer of split proteins, and is expressed during neuronal development.
Eur. J. Biochem. 258, 339349.
Greenman, Y., and Melmed, S. (1996). Diagnosis and management of nonfunctioning
pituitary tumors. Annu. Rev. Med. 47, 95106.
Hasson, P., and Paroush, Z. (2006). Crosstalk between the EGFR and other signalling
pathways at the level of the global transcriptional corepressor Groucho/TLE. Br. J. Cancer
94, 771775.
Horejsi, V., Drbal, K., Cebecauer, M., Cerny, J., Brdicka, T., Angelisova, P., and
Stockinger, H. (1999). GPI-microdomains: A role in signalling via immunoreceptors.
Immunol. Today 20, 356361.
Hough, C., Cho, K., Zonderman, A., Schwartz, D., and Morin, P. (2001). Coordinately
up-regulated genes in ovarian cancer. Cancer Res. 61, 38693876.
Ilangumaran, S., He, H., and Hoessli, D. (2000). Microdomains in lymphocyte signalling:
Beyond GPI-anchored proteins. Immunol. Today 21, 27.
Jhaveri, M., Rait, A., Chung, K., Trepel, J., and Chang, E. (2004). Antisense oligonucleo-
tides targeted to the human alpha folate receptor inhibit breast cancer cell growth and
sensitize the cells to doxorubicin treatment. Mol. Cancer Ther. 3, 15051512.
Kane, M. A., Portillo, R. M., Elwood, P. C., Antony, A. C., and Kolhouse, J. F. (1986). The
influence of extracellular folate concentration on methotrexate uptake by human KB
cells. Partial characterization of a membrane-associated methotrexate binding protein.
J. Biol. Chem. 261, 4449.
Katznelson, L., Alexander, J. M., and Klibanski, A. (1993). Clinical review 45: Clinically
nonfunctioning pituitary adenomas. J. Clin. Endocrinol. Metab. 76, 10891094.
Kelley, K., Rowan, B., and Ratnam, M. (2003). Modulation of the folate receptor alpha
gene by the estrogen receptor: Mechanism and implications in tumor targeting. Cancer
Res. 63, 28202828.
Luhrs, C. A., Raskin, C. A., Durbin, R., Wu, B., Sadasivan, E., McAllister, W., and
Rothenberg, S. P. (1992). Transfection of a glycosylated phosphatidylinositol-anchored
folate-binding protein complementary DNA provides cells with the ability to survive in
low folate medium. J. Clin. Invest. 90, 840847.
Mathias, C. J., and Green, M. A. (1998). A kit formulation for preparation of [(111)In]In-
DTPA-folate, a folate-receptor-targeted radiopharmaceutical. Nucl. Med. Biol. 25,
585587.
Mathias, C. J., Wang, S., Waters, D. J., Turek, J. J., Low, P. S., and Green, M. A. (1998).
Indium-111-DTPA-folate as a potential folate-receptor-targeted radiopharmaceutical.
J. Nucl. Med. 39, 15791585.
McCabe, C., Boelaert, K., Tannahill, L., Heaney, A., Stratford, A., Khaira, J., Hussain, S.,
Sheppard, M., Franklyn, J., and Gittoes, N. (2002). Vascular endothelial growth factor,
its receptor KDR/Flk-1, and pituitary tumor transforming gene in pituitary tumors.
J. Clin. Endocrinol. Metab. 87, 42384244.
McCabe, C., Khaira, J., Boelaert, K., Heaney, A., Tannahill, L., Hussain, S., Mitchell, R.,
Olliff, J., Sheppard, M., Franklyn, J., and Gittoes, N. (2003). Expression of pituitary
tumour transforming gene (PTTG) and fibroblast growth factor-2 (FGF-2) in human
pituitary adenomas: Relationships to clinical tumour behaviour. 1: Clin Endocrinol (Oxf ).
2003 Feb. 58, 141150.
Miotti, S., Canevari, S., Menard, S., Mezzanzanica, D., Porro, G., Pupa, S. M.,
Regazzoni, M., Tagliabue, E., and Colnaghi, M. I. (1987). Characterization of human
ovarian carcinoma-associated antigens defined by novel monoclonal antibodies with
tumor-restricted specificity. Int. J. Cancer 39, 297303.
Folate Receptor Expression in Pituitary Adenomas 265
Miotti, S., Bagnoli, M., Tomassetti, A., Colnaghi, M. I., and Canevari, S. (2000). Interaction
of folate receptor with signaling molecules lyn and G(alpha)(i-3) in detergent-resistant
complexes from the ovary carcinoma cell line IGROV1. J. Cell Sci. 113, 349357.
Moreno, C. S., Evans, C. O., Zhan, X., Okor, M., Desiderio, D. M., and Oyesiku, N. M.
(2005). Novel molecular signaling and classification of human clinically nonfunctional
pituitary adenomas identified by gene expression profiling and proteomic analyses. Cancer
Res. 65, 1021410222.
Onguru, O., Scheithauer, B., Kovacs, K., Vidal, S., Jin, L., Zhang, S., Ruebel, K., and
Lloyd, R. (2004). Analysis of epidermal growth factor receptor and activated epidermal
growth factor receptor expression in pituitary adenomas and carcinomas. Mod. Pathol. 17,
772780.
Orr, R. B., and Kamen, B. A. (1994). UMSCC38 cells amplified at 11q13 for the folate
receptor synthesize a mutant nonfunctional folate receptor. Cancer Res. 54, 39053911.
Orr, R. B., and Kamen, B. A. (1995). Identification of a point mutation in the folate receptor
gene that confers a dominant negative phenotype. Cancer Res. 55, 847852.
Rijnboutt, S., Jansen, G., Posthuma, G., Hynes, J. B., Schornagel, J. H., and Strous, G. J.
(1996). Endocytosis of GPI-linked membrane folate receptor-alpha. J. Cell Biol. 132,
3547.
Rochman, H., Selhub, J., and Karrison, T. (1985). Folate binding protein and the estrogen
receptor in breast cancer. Cancer Detect Prev. 8, 7175.
Ross, J. F., Chaudhuri, P. K., and Ratnam, M. (1994). Differential regulation of folate
receptor isoforms in normal and malignant tissues in vivo and in established cell lines.
Physiologic and clinical implications. Cancer 73, 24322443.
Semba, S., Han, S. Y., Ikeda, H., and Horii, A. (2001). Frequent nuclear accumulation of
beta-catenin in pituitary adenoma. Cancer 91, 4248.
Shen, F., Ross, J. F., Wang, X., and Ratnam, M. (1994). Identification of a novel folate
receptor, a truncated receptor, and receptor type beta in hematopoietic cells: cDNA
cloning, expression, immunoreactivity, and tissue specificity. Biochemistry 33, 12091215.
Sun, X. L., Murphy, B. R., Li, Q. J., Gullapalli, S., Mackins, J., Jayaram, H. N.,
Srivastava, A., and Antony, A. C. (1995). Transduction of folate receptor cDNA into
cervical carcinoma cells using recombinant adeno-associated virions delays cell prolifera-
tion in vitro and in vivo. J. Clin. Invest. 96, 15351547.
Sun, X. L., Jayaram, H. N., Gharehbaghi, K., Li, Q. J., Xiao, X., and Antony, A. C. (1999).
Modulation of the cytotoxicity of 3
0
-azido-3
0
-deoxythymidine and methotrexate after
transduction of folate receptor cDNA into human cervical carcinoma: Identification of a
correlation between folate receptor expression and thymidine kinase activity. Cancer Res.
59, 940946.
Toffoli, G., Cernigoi, C., Russo, A., Gallo, A., Bagnoli, M., and Boiocchi, M. (1997).
Over-expression of folate binding protein in ovarian cancers. Int. J. Cancer 74, 193198.
Wang, D., Zhang, H., Lang, F., and Yun, C. (2007). Acute activation of NHE3
by dexamethasone correlates with activation of SGK1 and requires a functional gluco-
corticoid receptor. Am. J. Physiol. Cell Physiol. 292, C396C404.
Yamada, S., Asa, S. L., Kovacs, K., Muller, P., and Smyth, H. S. (1989). Analysis of hormone
secretion by clinically nonfunctioning human pituitary adenomas using the reverse
hemolytic plaque assay. J. Clin. Endocrinol. Metab. 68, 7380.
266 Chheng-Orn Evans et al.
C H A P T E R N I N E
Regulation of Human Dihydrofolate
Reductase Activity and Expression
Emine Ercikan Abali, Nancy E. Skacel, Hilal Celikkaya, and
Yi-Ching Hsieh
Contents
I. Introduction 268
II. Structure and Binding of Dihydrofolate, MTX, and NADPH 269
III. Mechanism of DHFR Catalysis 272
IV. Alternative Substrates: Folic Acid and Dihydrobiopterin 273
V. Genomic Organization of DHFR 274
VI. Human Dihydrofolate Reductase Pseudogenes 276
VII. Transcriptional Regulation 276
VIII. Polymorphisms of DHFR 280
A. 19-bp deletion polymorphism in human DHFR intron-1 281
B. DHFR copy number variation in 9-bp tandem repeats 282
IX. Posttranscriptional Regulation of DHFR 283
X. Translational Regulation of DHFR 283
Acknowledgments 287
References 287
Abstract
Dihydrofolate reductase (DHFR) enzyme catalyzes tetrahydrofolate regenera-
tion by reduction of dihydrofolate using NADPH as a cofactor. Tetrahydrofolate
and its one carbon adducts are required for de novo synthesis of purines and
thymidylate, as well as glycine, methionine and serine. DHFR inhibition causes
disruption of purine and thymidylate biosynthesis and DNA replication, leading
to cell death. Therefore, DHFR has been an attractive target for chemotherapy of
many diseases including cancer. Over the following years, in order to develop
better antifolates, a detailed understanding of DHFR at every level has been
undertaken such as structure-functional analysis, mechanisms of action, tran-
scriptional and translation regulation of DHFR using a wide range of technolo-
gies. Because of this wealth of information created, DHFR has been used
Vitamins and Hormones, Volume 79
#
2008 Elsevier Inc.
ISSN 0083-6729, DOI: 10.1016/S0083-6729(08)00409-3 All rights reserved.
The Cancer Institute of New Jersey, Robert Wood Johnson Medical School, University of Medicine and
Dentistry of New Jersey, New Brunswick, New Jersey 08903
267
extensively as a model system for enzyme catalysis, investigating the relations
between structure in-silico structure-based drug design, transcription
from TATA-less promoters, regulation of transcription through the cell cycle,
and translational autoregulation. In this review, the current understanding of
human DHFR in terms of structure, function and regulation is summarized.
2008 Elsevier Inc.
I. Introduction
The enzyme dihydrofolate reductase (DHFR, 5,6,7,8-tetrahydrofolate:
NADPoxidoreductase, EC 1.51.3) catalyzes the reduction of dihydrofolate
(H
2
F) to tetrahydrofolate (H
4
F) utilizing NADPH as a cofactor. H
4
F and its
derivatives are essential cofactors in the synthesis of thymidylate, purines, and
some amino acids (Figs. 9.1Aand 9.2A) (Blakley and Cocco, 1984; Futterman,
1957; Osborn and Huennekens, 1958). Inhibition of DHFR results in a
depletion of the reduced folate pools, inhibition of DNA synthesis, and cell
death. Due to its biological significance, DHFRhas proven to be an important
target of antineoplastic, antiprotozoal, antifungal, and antimicrobial drugs in
addition to its use for the treatment of other nonmalignant diseases, such as
arthritis.
Antifolates are the oldest of the antimetabolite class of anticancer drugs
and have been used in the clinic for more than four decades. The first
clinically useful antifolate was aminopterin, a tight binding inhibitor of
DHFR. Treatment with aminopterin led to the first-ever remissions in
childhood leukemia. Soon after, methotrexate (MTX) replaced aminopterin
based on animal studies showing that MTX had a better therapeutic index.
Over the following years, in order to develop better antifolates, a detailed
understanding of DHFRat every level has been undertaken such as structure
functional analysis, mechanisms of action, transcriptional and translation regu-
lation of DHFR using a wide range of technologies. Because of this wealth of
information created, DHFR has been used extensively as a model system
for enzyme catalysis, investigating the relations between structure in silico
structure-based drug design, transcription from TATA-less promoters, regu-
lation of transcription through the cell cycle, and translational autoregulation.
In this review, the current understanding of human DHFR is summar-
ized. We begin with the structure and kinetic mechanism of enzyme of
DHFR. The review then concentrates on the genomic organization, poly-
morphisms, transcriptional, and translational regulation of DHFR. We refer
readers to earlier reviews that discuss many aspects of DHFR regulation for
additional in-depth analysis (Azizkhan et al., 1993; Banerjee et al., 2002;
Blakley, 1995; Cody and Schwalbe, 2006; Hammes-Schiffer and Benkovic,
2006; Schnell et al., 2004; Slansky and Farnham, 1996).
268 Emine Ercikan Abali et al.
II. Structure and Binding of Dihydrofolate,
MTX, and NADPH
As a result of its importance, a detailed picture how DHFR works
emerged by studying its kinetics and structure; however, the most extensive
structural characterization has been done for Escherichia coli. Fortunately, the
amino acids required for catalysis and the general features of the secondary
NH
2
H
2
N
4
5
6
9
7
8
2
1
N
N
N
N
N
O
O OH
OH
O
NH
Pteridine pABA
pABG
Methotrexate (MTX)
4
2
1
5
8
7
6
9
10
OH
H
2
N
N
N N
H
NH
NH
O
O OH
OH
O
N
Dihydrofolate (H
2
F)
Glu
4
5
6 8 2
H
2
N
O
O
O
O
O P P
P
O
O
O
O
O
OH
O
HO OH HO
N
N
N
N N
NH
2
Nicotinamide
Adenine
NADPH
Phosphoribose
A
B
NADPH
Antifolate
Loop I
B
g
h
f
a
e
b
c d
F
C
E
Figure 9.1 (A) Structures and atomic numbering of methotrexate (MTX), dihydro-
folate (H
2
F), and NADPH. (B) a-Carbon representation of human DHFR complexed
with NADPH and MOT, a furo analog of folate (N-[(4-phenyl)carbonyl]-L-glutamic
acid). b-Sheets are labeled with lower case letters and a-helices are labeled with upper
case letters (PDB code 1hfq) (Cody et al., 1998).
Human Dihydrofolate Reductase 269
structures have been conserved throughout the evolution expect in protozoa
and plants where DHFR is fused to thymidylate synthase as a bifunctional
enzyme.
DHFR shares a modified version of a common motif that is highly
conserved within the enzymes that bind NAD(H) or NADP(H). This
motif called the dinucleotide-binding domain, or Rossmann fold consists
of an open, six-stranded, parallel b-sheet flanked on each side by a-helices
(Bottoms et al., 2002; Carugo and Argos, 1997a,b; Rossmann et al., 1974).
In the case of DHFR, several parallel and antiparallel b-strands are inter-
calated by a-helices (Carugo and Argos, 1997b).
Human DHFR is a monomeric protein of 186 amino acids with a
molecular weight of 21,544 Da (Blakley, 1995). The core of the human
DHFR has an eight-stranded b-sheet consisting of seven parallel strands and
a carboxy-terminal antiparallel strand. The seventh strand consists of two
consecutive strands, bG1 and bG2 due to a seven-residue disruption
between residues 160 and 167, resulting in a tight turn. Five a-helices are
HN
N
2
2
2
2
1 1
5
5
8 8
7 7
6
6
5
5
6
6
4
4
4
4
N
H
H
2
N
N
H
N
H
2
N
H
2
N
HN H
O
O
N
N
H
N
N
H
O
R
1
R
2
R
2
R
2
H
2
N
A
O
B
H
+
B
E
NADPH
H
2
F
E
NADP
+
H
4
F
E
NADPH
H
4
F
E
H
4
F
E
NADPH
225s
1
84s
1
14mM
1
s
1
0.7mM
1
s
1
4.4mM
1
s
1
100s
1
94s
1
98mM
1
s
1
1360s
1
37s
1
Figure 9.2 (A) Reduction of H
2
F by DHFR. (B) Kinetic scheme of the catalytic cycle
of DHFR denoted as E. The five kinetic intermediates and the rate constants at pH 7.65
at 20
C are shown (Blakley, 1995).
270 Emine Ercikan Abali et al.
packed against the b-sheet core (Fig. 9.1B). The helix aE
0
, which is imme-
diately after aE is perpendicular to aE and may have emerged due to five
residue insertion in human DHFR relative to bacterial enzymes (Davies
et al., 1990). In addition, vertebrate including human DHFR has one left-
handed, type II polyproline-like helix, which is not present in E. coli.
Another deviation fromE. coli DHFRis the presence of a cis-peptide linkage,
one between residues Arg65 and Pro66. The other cis-peptide linkage is a
conserved structural feature of all DHFRs and is between residues Gly116
and Gly117, which are near the nicotinamide-binding site (Blakley, 1995;
Cody and Schwalbe, 2006; Davies et al., 1990). The active site cleft is formed
at the junction of two subdomains: a larger subdomain that binds the
adenosine portion of NADPH and the smaller loop domain. The active
site is enclosed in a large hydrophobic pocket surrounded by the a-helix B,
the central b-sheets (a, e, and b) and the loop 1. Two polar groups, Glu30 and
Arg70 at each end, seal this hydrophobic pocket. The acidic residue, Glu30
interacts with the 2-amino group and N3 of the pteridine ring folate and
the 2-amino group and N1 of MTX. The basic group, Arg70 interacts
with the a-carboxylate of the glutamate moiety of both the substrate and
the inhibitor. The nicotinamide moiety also resides in this deep hydrophobic
pocket in the vicinity of the substrate, but the rest of the NADPHbinds in an
extended conformation with the 2
0
-phosphoADP-ribose moiety occupying
a cleft which includes the carboxy-terminal ends of five strands of b-sheet
(Cody et al., 1992).
The conformation of vertebrate DHFRs including human DHFR is
more rigid than E. coli DHFR. However, the amino acids required for
catalysis and the general features of the secondary structures, as well as the
kinetic pathways are all conserved throughout the evolution with the
exception of the flexibility of the enzymes between vertebrates and E. coli.
Comparison of crystal structures of DHFR reveals that unlike E. coli
DHFR, vertebrate DHFRs lack loop 1 motion and subdomain rotation
(Met20 loop in E.coli) structures (Sawaya and Kraut, 1997). While there are
no obvious structural explanation for the lack of subdomain rotation in
vertebrates, there are three structural differences between vertebrate and
E. coli DHFR that may give rise to the rigidity of vertebrate DHFRs:
insertion of a left-handed polyproline-type helix in vertebrate DHFRs
loop 1, Gly20 in the loop 1 of vertebrate DHFR instead of an asparagine
which creates a stable b-hairpin, and the truncation of the vertebrates GH
loop preventing the formation of hydrogen bonds with loop 1.
The structures of folate and MTX are essentially the same except for the
4-amino group and a methyl group attached to N10 in MTX instead of a
4-oxo group and hydrogen attached to N10 of folate (Fig. 9.1A). However,
MTX binding to human DHFR is quite different than that of folate,
consequently making different interactions with the active site residues.
MTX is protonated at the N1 position of the pteridine ring and is also
Human Dihydrofolate Reductase 271
rotated 180
N
O
O
O
P
SH
SAM
CH
3
HO
HO
HO
N
N
N
NH
2
NH
2
NH
2
O
O
O
O
O
O
OH
O O
P
O
O
O
N
H
H
H
N
N
N
N
N
HO
NH
O
Co
N
N
N
N
H
H
H
MTHF
CH
3
CH
3
CH
3
-B
12
-OH
CH
3
-OAryl
CH
3
-Cl
CH
3
-NH
2
CH
3
-SR
CH
3
-SCoM (methyl-SCoM)
CH
3
-H
4
folate (MTHF)
CH
3
-H
4
MPT
CH
3
-
+
SR (SAM)
CH
3
-Co (CH
3
-B
12
)
CH
3
-Ni
H
COOH
COOH
H
O
O
N
N
Figure 10.1 Methyl donors in biology.
294 Stephen W. Ragsdale
that are found in nature like methanol, methylamines, methanethiols as well
as the cofactors that are involved in methylation reactions like methyltetra-
hydrofolate (MTHF), S-adenosyl-L -methionine (SAM), and methyl-B
12
.
A methyltransferase catalyzes the transfer of a methyl group from one of
the donors shown in Fig. 10.1, like MTHF or SAM, to an acceptor like
homocysteine or the N6 group on adenine in DNA (Fig. 10.2).
Two cofactors figure prominently in methyltransferase chemistry: vitamin
B
12
(or cobalamin) and tetrahydrofolate (THF). The history of vitamin B
12
dates back to its description as the antipernicious anemia factor (Minot and
Murphy, 1926; Whipple and Robscheit-Robbins, 1925). Cobalamin was
isolated by Smith and Folkers in 1948 (Rickes et al., 1948; Smith, 1948) and
its structure was crystallized (Rickes et al., 1948) and its structure was deter-
mined in 1956 (Hodgkin et al., 1956). Like heme, F
430
, and chlorophyll,
vitamin B
12
is a tetrapyrrolic cofactor with a central Co atom coordinated by
the four equatorial pyrrole nitrogen ligands (Fig. 10.1). Extending fromone of
the pyrrole rings is a propanolamine-linked group that can serve as the lower
axial ligand to Co. The lower ligand is dimethylbenzimidazole in cobalamins,
while the upper axial ligand is a cyano-, methyl-, or 5
0
-deoxyadenosyl- group
in vitamin B
12
, methylcobalamin (MeCbl
1
), and adenosylcobalamin(AdoCbl)
or coenzyme B
12
, respectively. The biological role of MeCbl as an essential
coenzyme for a methyltransferase (Guest et al., 1962) was revealed a few years
after a role for AdoCbl as the coenzyme for glutamate mutase was discovered
(Barker et al., 1958).
Tetrahydrofolate (THF) is the reduced form of the vitamin folic acid,
which was first recognized as the compound in brewers yeast that can
reverse anemia. Folic acid was isolated in a highly purified form from 4 tons
of spinach leaves by Esmond Snell, Herschel Mitchell, and Roger Williams
(Mitchell et al., 1941). The vitamin was crystallized soon after (Pfiffner et al.,
1945; Stokstad, 1943) and synthesized in 1946 by Lederle Laboratories
(Angier et al., 1946). Besides serving as a cofactor in methyltransferase
reactions, THF is the major one-carbon carrier in cells and is needed for
protein and DNA synthesis and is important in nitrogen metabolism. Its
essentiality in such key metabolic processes makes THF metabolism a key
target for anticancer and antimicrobial drugs (M.P. Costi, S. Ferrari, 2001).
CH
3
-H
4
folate
CH
3
-
+
SR
SAM
MTHF
Methyltransferases
SR
Y (DNA, histone, serotonin, dopamine)
RS
(Homocysteine)
CH
3
-Y
CH
3
-Co(III)
CH
3
-RS (Methionine) Co(I)
H
4
folate
Figure 10.2 Two types of methyltransferases.
Tetrahydrofolate and B
12
in Methyltransferases 295
There are two classes of methyltransferases, which differ in their use of
an activated versus an unactivated methyl group donor. One class of methyl-
transferases, exemplified by methionine synthase, use an unactivated methyl
donor and an intermediate methyl carrier, cobalamin, which in the Co(I)
oxidation state is an extremely potent nucleophile that can react with the
methyl group of MTHF to generate an intermediate organometallic MeCbl
intermediate. The methyl group is then transferred from MeCbl to the final
acceptor, which for methionine synthase is homocysteine, generating
methionine. The other class of methyltransferases uses an activated methyl
donor in the form of SAM to directly methylate the N6 of adenine or the
amino groups of lysine or arginine in histones, and so on. This review
focuses on the first class of methyltransferases that utilize THF and B
12
.
In all methyltransferase-catalyzed reactions, as shown in Fig. 10.3, trans-
fer of the methyl group involves cleavage of a methyl-X bond, where X
can be one of a variety of functional groups (N, S, Cl, O, etc.). Heterolysis
can lead to two products: a methyl cation or a methyl anion. On the contrary,
homolysis leads to a methyl radical. For methyl group transfers, heterolysis is
most common, leading to the formation of a methyl carbocation equivalent.
II. Three Component Systems Required for
B
12
/THF-Dependent Methyltransferases
All B
12
-dependent methyltransferases contain three components, as
indicated in Fig. 10.4 and Table 10.1. In these systems, the B component
binds the methyl donor, a C component binds B
12
, and an A compo-
nent binds the methyl group acceptor. An alternative nomenclature in use
describes the B component as MT1 and the A component as MT2. Recent
literature references for each methyltransferase system are given in Table 10.1.
In each case, the methyl group is unactivated, that is, it is a secondary alcohol,
an amine (primary, secondary, or tertiary), or a thiolate. The required
electrophilic activation is accomplished by the B component, as described
in more detail below. The C component supplies the supernuclophilic
Co(I) state of B
12
, which acts as an intermediary methyl group acceptor that
interacts with the B component to accept the methyl group from the
Intermediates
CH
3
-X
[ CH
3
+
+X
] Heterolysis
[ CH
3
+X
+
] Heterolysis
[ CH
3
+X
] Homolysis
Figure 10.3 Three ways to cleave a methyl-X bond.
296 Stephen W. Ragsdale
substrate, forming MeCbl. The C Component then interacts with the A
component to transfer the methyl group to the ultimate methyl group
acceptor.
III. Biological Systems Impacted by B
12
and
Folate-Dependent Methyltransferases
A. Methionine biosynthesis
B
12
-dependent methyltransferases play an important role in microbial and
eukaryotic metabolism. Biosynthesis of methionine in most microbes and
eukaryotes depends upon the B
12
- and THF-dependent enzyme, methio-
nine synthase (MetH). Since methionine is the precursor of SAM, methio-
nine synthase supports the many methyltransferases involved in methylation
of DNA, proteins, and neurotransmittors (Fig. 10.2). An elevated level of
homocysteine is linked to a number of pathological states, including prema-
ture heart disease and neural tube defects and methionine synthase plays an
important role in controlling homocysteine homeostasis (Col et al., 2007).
B. Methanogenesis
As shown in Fig. 10.5, growth of methanogens requires the conversion of
substrates (CO
2
, methylamines, methylthiols, and acetate) to methyl-
SCoM. The nickel containing enzyme methyl-SCoM reductase then
catalyzes the conversion of methyl-SCoM to methane. Table 10.1 lists six
B
12
-dependent methyltransferases that are important in growth of metha-
nogenic archaea. Most of these methyltransferases catalyze the transfer of
a methyl group from the methyl donor (Methyl-X) to CoM, thus providing
methyl-SCoM needed for methane formation. These methanogenic
methyltransferases use B
12
(or an analog with a slight modification in the
MtxB,
CmuB,
MetH,
AcsE,
MtrH
MtxC,
CmuA,
MetH,
AcsCD,
CdhDE
MtxA,
CmuA,
MetH,
AcsAB,
CdhABC
MT1
or
B
MT2
or
A
X=
OH, OR,
OAr, NR
2,
SH, Cl,
THF,
THMPT
X
CH
3
-X
Y
Y=
HCys, CoM,
CODH/ACS,
ACDS
CH
3
-Y
CH
3
Co
I
Co
III
Figure 10.4 B
12
-dependent methyltransferases.
Tetrahydrofolate and B
12
in Methyltransferases 297
Table 10.1 B
12
-dependent methyltransferases
Methyl donor
Methyltransferase
components Biological system
Ultimate methyl
acceptor Reference
MTHF MetH Methionine
Synthesis
Homocysteine (Matthews, 2001)
CH
3
NH MtmABC Methanogenesis Coenzyme M (Krzycki, 2004)
(CH
3
)
2
N MtbABC Methanogenesis Coenzyme M (Soares et al., 2005)
(CH
3
)
3
N MttABC Methanogenesis Coenzyme M (Soares et al., 2005)
CH
3
SH MtsAB Methanogenesis Coenzyme M (Tallant et al., 2001)
MTHMPT MtrA-H Methanogenesis CoM (Gottschalk and Thauer, 2001)
CH
3
OH MtaABC Methanogenesis Coenzyme M (Hagemeier et al., 2006)
MTHF AcsABCDE Acetogenesis THF, CODH/ACS (Doukov et al., 2007)
MTHF MtaABC Acetogenesis THF (Das et al., 2007)
CH
3
OAr MtvABC Acetogenesis THF (Naidu and Ragsdale, 2001)
CH
3
Cl CmuAB Dehalorespiration THF (Studer et al., 2001)
benzimidazole component) bound to the C component. Growth on
acetate or H
2
/CO
2
involves the generation of methyltetrahydromethanop-
terin (MTHMPT), which is converted to methyl-SCoM by the Mtr system.
As described below, this is interesting membrane-bound system links the
methyl transfer reaction to generation of a sodium ion gradient, which is
utilized to make ATP (Gottschalk and Thauer, 2001).
C. Methanogenic methyltransferases
and the 22nd amino acid
Study of the methanogenic methyltransferases has uncovered the 22nd
amino acid, pyrrolysine, which is encoded by a UAG stop codon. This
system is reminiscent of selenocysteine, encoded by the UGA codon. For
pyrrolysine synthesis, there is an enzymatic system, encoded by the pylBCD
genes to synthesize pyrrolysine, a dedicated pyrrolysyl-tRNA synthetase
(encoded by pylS), which aminoacylates a specific amber-decoding tRNA
(encoded by the pylT gene) with pyrrolysine (Blight et al., 2004; Longstaff
et al., 2007). Thus, like the other 20 natural amino acids, pyrrolysine is
co-translationally placed into the nascent polypeptide chain. The role of
pyrrolysine in the methyltransferase reaction is discussed below.
D. Acetogenesis
A role for B
12
in anaerobic CO
2
fixation was described in 1965 (Ljungdahl
et al., 1966; Poston et al., 1964). There are two B
12
-dependent methyl-
transferases involved in the Wood-Ljungdahl pathway of CO
2
fixation
(Fig. 10.6). The first, MTHF: corrinoid iron-sulfur protein (CFeSP)
methyltransferase is encoded by the acsE gene and catalyzes transfer of the
methyl group from MTHF to the Co(I)-B
12
site in the CFeSP, which is
encoded by the acsCD genes. In the subsequent reaction, the methylated
Methyl-Coenzyme M Reductase
MCR
CH
4
CH
3
-SCoM
CH
3
COOH
CH
3
-NR
2
(R=H, CH
3
)
CH
3
-SH
CH
3
-OH
MeTr
MeTr
MeTr
CO
2
N N
N
N
O
O
N
Nl(l)
A
B
C
D
Figure 10.5 Methanogenesis.
Tetrahydrofolate and B
12
in Methyltransferases 299
CFeSP transfers the methyl group to a NiFeS cluster in acetyl-CoA synthase
(ACS); thus, this is a metal (Co) to metal (Ni) methyl group transfer between
two proteins. ACS then catalyzes the condensation of its Ni-bound methyl
group with CO (generated by CO dehydrogenase, CODH) and coenzyme
A to generate acetyl-CoA. The series of organometallic intermediates is one
of the novel features of the Wood-Ljungdahl pathway.
Two other methyltransferase systems that are shown in Fig. 10.6 couple
to the Wood-Ljungdahl pathway. The MtvABCsystem transfers the methyl
group of vanillate or other methoxylated aromatics to THF to gener-
ate MTHF (Engelmann et al., 2001; Naidu and Ragsdale, 2001), while
MtaABC catalyzes the synthesis of MTHF from the methyl group of
methanol and THF (Das et al., 2007). MtvB was shown to catalyze methyl
transfer from the aromatic phenylmethyl ether to the corrinoid component
of MtvC and MtvA catalyzed the MtvC-dependent methylation of THF
(Naidu and Ragsdale, 2001). The sequence homology between the Mtv
and Mta systems suggests that the A, B, and C components in the two
systems share similar functions (Das et al., 2007).
E. Other metabolic systems in which B
12
- and
folate-dependent methyltransferases play a key role
Growth of Sphingomonas paucimobilis SYK-6 on lignin-derived biaryls and
monomers requires the ligM gene, which is homologous to mtvB, to catalyze
the transfer of the methyl group from vanillate to THF. Since the ligM gene
in this organism is located in the same gene cluster as two genes encoding
THF-dependent enzymes, it is likely that lignin degradation is linked
directly to one-carbon metabolism (Abe et al., 2005). However, unlike
the methyltransferase reactions described in the previous section, these
remain to be characterized.
CO
2
HCOOH
MeTr
MtvABC or
MtaABC
H
3
C
O
C
CO
CODH
H
2
H
2
H
2
H
2
CO
2
ACS
Co(I)
CFeSP
CH
3
-Co(III)
CH
3
-THF
CH
3
-OAr or
CH
3
-OH
THF
THF
CoA
SCoA
Figure 10.6 Methyltransferases in acetogenesis.
300 Stephen W. Ragsdale
IV. Structure and Function of B
12
in Methyltransferases
A. Binding of B
12
to the enzymes
Three modes of B
12
binding have been described: dmb-on, dmb-off /
his-on, and base-off, which refers to whether or not a lower axial
ligand is coordinated to Co (Fig. 10.7). Thus, dmb-on indicates that the
dimethylbenzimidazole group, which is appended to one of the tetrapyrrole
rings, is ligated to Co, while his-on refers to the state in which the dmb
ligand is replaced by a His residue donated by the protein. Dmb-on and
dmb-off have often been referred to as base-off and base-on;
however, this is inaccurate because the His ligand also acts as a base. It is
more accurate to only refer to the base-off binding mode when there is
no nitrogenous base ligand. Such a coordination mode was first revealed by
spectroscopic studies of the CFeSP in the Wood-Ljungdahl pathway
(Ragsdale et al., 1987) and recently confirmed in the crystal structure
(Svetlitchnaia et al., 2006). In most cases, unambiguous definition of the
ligation state, that is, dmb-on, dmb-off , base-off , was first revealed
by spectroscopic studies. For example, a dmb-off state was revealed for
several proteins by electron paramagnetic resonance (EPR) spectroscopy
(Ragsdale et al., 1987) (Stupperich, 1990, #220) several years before a crystal
structure revealed a his-ondmb-off binding mode in the B
12
-binding
domain of methionine synthase (Drennan et al., 1994). EPR spectroscopic
studies have been successful in revealing the ligation mode in a number of
other enzymes (Abend et al., 1998, #6281; Yamanishi et al., 1998, #2310;
Lawrence et al., 1999, #4157; Abend et al., 1999, #3897; Ke et al., 1999,
#6277). The EPR spectrum of the Co(II) state of B
12
is particularly
diagnostic of the axial ligation state because the unpaired electron in the
d
2
Z
orbital exhibits strong interactions with the axial ligand. If this ligand is a
nitrogen atom, with a nuclear spin (I ) of 1, each of the eight hyperfine lines
(due to splitting of the resonance by interactions with the Co nucleus with
I 7/2) in the EPR spectrum exhibit a three-line splitting. The base-off
mode of B
12
binding is characterized by the presence of singlets instead of
Figure 10.7 Three modes of B
12
binding.
Tetrahydrofolate and B
12
in Methyltransferases 301
triplets at each of the eight resonant positions, as was observed in the EPR
spectrum of the CFeSP (Ragsdale et al., 1987). The his-on mode is
clearly indicated by adding
15
N-His to the growth medium, which replaces
the natural abundance histidine (mostly
14
N) in enzymes. Since
15
N has a
nuclear spin of , doublet instead of triplet superhyperfine lines are
observed at each of the eight resonant positions in the EPR spectrum.
Observation of the triplet spectrum in B
12
proteins labeled with
15
N-His
suggests a dmb-on binding mode.
Crystallographic studies of methyltransferases have revealed the elegant
molecular details of how proteins bind B
12
. The crystal structures of the
cobalamin-binding component of several methyltransferases are available,
including the B
12
-binding domain of methionine synthase in several states
(Bandarian et al., 2002, 2003; Drennan et al., 1994), MtaC (Das et al., 2007 )
(Fig. 10.8) and the CFeSP (AcsCD) from M. thermoacetica (Svetlitchnaia
et al., 2006) (Fig. 10.9), and the MtaBC complex from Methanosarcina barkeri
(Hagemeier et al., 2006). In all of these methyltransferase structures, cobala-
min is bound within a Rossman a/b fold (Fig. 10.10). The C compo-
nents that have a dmb-off/his-on mode of B
12
binding share a sequence
motif DXHXXGX
41
SXLX
2628
GG in which the His is the lower axial
ligand. The Asp, His, and Ser residues in this sequence are referred to as
the catalytic triad and facilitate formation of the base-off conformation
by protonating the His ligand, (Ludwig and Matthews, 1997). In the
M. thermoacetica MtaC, the Asp and His are present in the catalytic triad
(Fig. 10.6); however, it appears that a Thr residue replaces Ser. The dmb
Figure 10.8 B
12
-binding site in M. thermoacetica MtaC (Das et al., 2007). The region
containing the conserved DXH motif (see text) is shown in cornflower blue. Generated
from PDB ID# 1Y80 using Chimera.
302 Stephen W. Ragsdale
side chain is deeply embedded and is responsible for much of the binding
energy that tethers the cobalamin to the protein. A dmb-on structure is
not available in the methyltransferase class, which is somewhat surprising since
several AdoCbl-dependent isomerases share the dmb-on binding mode,
Figure 10.9 B
12
-binding site in the M. thermoacetica CFeSP (Svetlitchnaia et al., 2006 ).
The region containing the hydrophobic helix (see text) is shown in cornflower blue.
Generated using Chimera from PDB ID code 2H9A.
Figure 10.10 B
12
-binding site in M. barkeri MtaC focusing on the Rossman domain
involved in ligating the Dmb moiety, generated from PDB ID code 2I2X, using
Chimera.
Tetrahydrofolate and B
12
in Methyltransferases 303
including diol dehydratase (Abend et al., 1998; Shibata et al., 1999; Yamanishi
et al., 1998), ribonucleotide reductase (Lawrence et al., 1999; Sintchak et al.,
2002), and ethanolamine ammonia lyase (Abend et al., 1999; Ke et al., 1999).
Most of the methyltransferases share the dmb-off/his-on binding
mode, while the CFeSP involved in transferring the methyl group to the
ACS component in the Wood-Ljungdahl pathway is in the base-off
state, as revealed by spectroscopic studies (Ragsdale et al., 1987) and con-
firmed by crystallographic studies (Svetlitchnaia et al., 2006) of the CFeSP
(Fig. 10.9). A related protein in the methanogenic acetyl-CoA decarbony-
lase synthase complex also is in the base-off state ( Jablonski et al., 1993).
Proteins that bind B
12
in the base-off state lack the DXH . . . signature
sequence. Instead, they contain a relatively hydrophobic helix below the
plane of the cobalamin (SVLTAWAA) (Fig. 10.9). A coordinating water
molecule in the upper axial position is replaced by a methyl group during
catalysis. The EPR spectrum of the CFeSP in H
2
17
O exhibits
17
O-induced
hyperfine broadening, providing conclusive demonstration that H
2
O
coordinates to the metal center in one of the open axial positions (Stich
et al., 2006).
B. Generation and maintenance of the active Co(I) state of B
12
Cobalt in B
12
can exist in the (I), (II), and (III) states. Cobalt cycles between
the Co(I) and methyl-Co(III) states during catalysis. In the Co(I) state, Co
has a d
8
configuration. In B
12
and related corrinoids, Co(I) is a super-
nucleophile (Schrauzer and Deutsch, 1969; Schrauzer et al., 1968) and is
weakly basic, with a pK
a
below 1 for the Co(I)-H complex (Tackett et al.,
1963). Protein-bound Co(I) is also highly reducing with a standard reduc-
tion potential for the Co(II)/(I) couple below 500 mV (Banerjee et al.,
1990b). These properties make Co(I) fairly unstable and subject to inacti-
vation. For example, in cobalamin-independent methionine synthase, the
Co(I) center undergoes oxidative inactivation to the 2
X-ray structure of B
12
-binding domains of methionine
synthase. Science 266, 16691674.
Drummond, J. T., Huang, S., Blumenthal, R. M., and Matthews, R. G. (1993). Assignment
of enzymatic function to specific protein regions of cobalamin-dependent methionine
synthase from Escherichia coli. Biochem. 32, 92909295.
Engelmann, T., Kaufmann, F., and Diekert, G. (2001). Isolation and characterization of a
veratrol:corrinoid protein methyl transferase from Acetobacterium dehalogenans. Arch.
Microbiol. 175, 376383.
Ensign, S. A., and Allen, J. R. (2003). Aliphatic epoxide carboxylation. Annu. Rev. Biochem.
72, 5576.
Evans, J. C., Huddler, D. P., Hilgers, M. T., Romanchuk, G., Matthews, R. G., and
Ludwig, M. L. (2004). Structures of the N-terminal modules imply large domain motions
during catalysis by methionine synthase. Proc. Natl. Acad. Sci. USA 101, 37293736.
Evans, J. C., Huddler, D. P., Jiracek, J., Castro, C., Millian, N. S., Garrow, T. A., and
Ludwig, M. L. (2002). Betaine-homocysteine methyltransferase: zinc in a distorted
barrel. Structure 10, 11591171.
Fasching, M., Schmidt, W., Krautler, B., Stupperich, E., Schmidt, A., and Kratky, C.
(2000). Co alpha-(1H-imidazolyl)-Co beta-methylcob(III)amide: Model for protein-
bound corrinoid cofactors. Helv. Chim. Acta. 83, 22952316.
Fedorov, A., Shi, W., Kicska, G., Fedorov, E., Tyler, P. C., Furneaux, R. H., Hanson, J. C.,
Gainsford, G. J., Larese, J. Z., Schramm, V. L., and Almo, S. C. (2001). Transition
state structure of purine nucleoside phosphorylase and principles of atomic motion in
enzymatic catalysis. Biochemistry 40, 853860.
Fujii, K., Galivan, J. H., and Huennekens, F. M. (1977). Activation of methionine synthase:
further characterization of flavoprotein system. Arch. Biochem. Biophys. 178, 662670.
Gencic, S., LeClerc, G. M., Gorlatova, N., Peariso, K., Penner-Hahn, J. E., and
Grahame, D. A. (2001). Zinc-thiolate intermediate in catalysis of methyl group transfer
in Methanosarcina barkeri. Biochemistry 40, 1306813078.
Gottschalk, G., and Thauer, R. K. (2001). The Na()-translocating methyltransferase
complex from methanogenic archaea. Biochim. Biophys. Acta. 1505, 2836.
Goulding, C. W., and Matthews, R. G. (1997). Cobalamin-dependent methionine synthase
from Escherichia coli: involvement of zinc in homocysteine activation. Biochemistry 36,
1574915757.
Guest, J. R., Friedman, S., Woods, D. D., and Smith, E. L. (1962). A methyl analog of
cobamide coenzyme in relation to methionine synthesis by bacteria. Nature 195, 340342.
Tetrahydrofolate and B
12
in Methyltransferases 319
Hagemeier, C. H., Krer, M., Thauer, R. K., Warkentin, E., and Ermler, U. (2006). Insight
into the mechanism of biological methanol activation based on the crystal structure of the
methanol-cobalamin methyltransferase complex. Proc. Natl. Acad. Sci. USA 103,
1891718922.
Hampele, I. C., DArcy, A., Dale, G. E., Kostrewa, D., Nielsen, J., Oefner, C., Page, M. G.,
Schonfeld, H. J., Stuber, D., and Then, R. L. (1997). Structure and function of the
dihydropteroate synthase from Staphylococcus aureus. J. Mol. Biol. 268, 2130.
Hao, B., Gong, W., Ferguson, T. K., James, C. M., Krzycki, J. A., and Chan, M. K. (2002).
A new UAG-encoded residue in the structure of a methanogen methyltransferase. Science
296, 14621466.
Hao, B., Zhao, G., Kang, P. T., Soares, J. A., Ferguson, T. K., Gallucci, J., Krzycki, J. A.,
and Chan, M. K. (2004). Reactivity and chemical synthesis of L-pyrrolysine- the 22(nd)
genetically encoded amino acid. Chem. Biol. 11, 13171324.
Harder, S. A., Lu, W.-P., Feinberg, B. F., and Ragsdale, S. W. (1989). Spectroelectrochem-
ical studies of the corrinoid/iron-sulfur protein from Clostridium thermoaceticum.
Biochemistry 28, 90809087.
Hilhorst, E., Iskander, A. S., Chen, T. B. R. A., and Pandit, U. (1994). Alkyl transfer from
quaternary ammonium salts to cobalt(I): model for the cobalamin-dependent methionine
synthase reaction. Tetrahedron 50, 88638870.
Hodgkin, D. C., Kamper, J., Mackay, M., Pickworth, J., Trueblood, K. N., and White, J. G.
(1956). Structure of vitamin B12. Nature 178, 6466.
Hogenkamp, H. P., Rush, J. E., and Swenson, C. A. (1965). Observations on the organo-
metallic bond of the corrinoid coenzymes. J. Biol. Chem. 240, 36413644.
Hoover, D. M., Jarrett, J. T., Sands, R. H., Dunham, W. R., Ludwig, M. L., and
Matthews, R. G. (1997). Interaction of Escherichia coli cobalamin-dependent methio-
nine synthase and its physiological partner flavodoxin: Binding of flavodoxin leads to axial
ligand dissociation from the cobalamin cofactor. Biochemistry 36, 127138.
Huang, C. C., Casey, P. J., and Fierke, C. A. (1997). Evidence for a catalytic role of zinc in
protein farnesyltransferase. Spectroscopy of Co2-farnesyltransferase indicates metal
coordination of the substrate thiolate. J. Biol. Chem. 272, 2023.
Jablonski, P. E., Lu, W. P., Ragsdale, S. W., and Ferry, J. G. (1993). Characterization of the
metal centers of the corrinoid/iron-sulfur component of the CO dehydrogenase enzyme
complex from Methanosarcina thermophila by EPR spectroscopy and spectroelectro-
chemistry. J. Biol. Chem. 268, 325329.
Kaufmann, F., Wohlfarth, G., and Diekert, G. (1998). O-demethylase from Acetobacterium
dehalogenans--substrate specificity and function of the participating proteins. Eur. J.
Biochem. 253, 706711.
Ke, S. C., Torrent, M., Museav, D. G., Morokuma, K., and Warncke, K. (1999). Identifi-
cation of dimethylbenzimidazole axial coordination and characterization of (14)N super-
hyperfine and nuclear quadrupole coupling in Cob(II)alamin bound to ethanolamine
deaminase in a catalytically-engaged substrate radical-Cobalt(II) biradical state. Biochemistry
38, 1268112689.
Kicska, G. A., Tyler, P. C., Evans, G. B., Furneaux, R. H., Shi, W., Fedorov, A.,
Lewandowicz, A., Cahill, S. M., Almo, S. C., and Schramm, V. L. (2002). Atomic
dissection of the hydrogen bond network for transition-state analogue binding to purine
nucleoside phosphorylase. Biochemistry 41, 1448914498.
Krautler, B. (1987). Thermodynamic trans-effects of the nucleotide base in the B
12
coenzymes. Helvetica Chimica Acta. 70, 12681278.
Kruer, M., Haumann, M., Meyer-Klaucke, W., Thauer, R. K., and Dau, H. (2002). The
role of zinc in the methylation of the coenzyme M thiol group in methanol:coenzyme M
methyltransferase from Methanosarcina barkeri. Eur. J. Biochem. 269, 21172123.
320 Stephen W. Ragsdale
Krzycki, J. A. (2004). Function of genetically encoded pyrrolysine in corrinoid-dependent
methylamine methyltransferases. Curr. Opin. Chem. Biol. 8, 484491.
Lane, T. J., and Quinlan, K. P. (1960). Metal binding of the benzimidazoles. J. Am. Chem.
Soc. 82, 29942997.
Lawrence, C. C., Gerfen, G. J., Samano, V., Nitsche, R., Robins, M. J., Retey, J., and
Stubbe, J. (1999). Binding of Cob(II)alamin to the adenosylcobalamin-dependent ribo-
nucleotide reductase from Lactobacillus leichmannii. Identification of dimethylbenzimi-
dazole as the axial ligand. J. Biol. Chem. 274, 70397042.
Lebertz, H., Simon, H., Courtney, L. F., Benkovic, S. J., Zydowsky, L. D., Lee, K., and
Floss, H. G. (1987). Stereochemistry of acetic acid formation from 5-methyl-tetrahy-
drofolate by Clostridium thermoaceticum. J. Am. Chem. Soc. 109, 31733174.
Lexa, D., and Saveant, J.-M. (1976). Electrochemistry of vitamin B
12
. I. Role of the base-
on/base-off reaction in the oxidoreduction mechanism of the B
12r
-B
12s
system. J. Am.
Chem. Soc. 98, 26522658.
Lindahl, P. A. (2004). Acetyl-Coenzyme A Synthase: The Case for a Ni
p
0
-Based Mechanism
of Catalysis. J. Biol. Inorg. Chem. 9, 516524.
Ljungdahl, L., Irion, E., and Wood, H. G. (1966). Role of corrinoids in the total synthesis of
acetate from CO
2
by Clostridium thermoaceticum. Fed. Proc. 25, 16421648.
Longstaff, D. G., Larue, R. C., Faust, J. E., Mahapatra, A., Zhang, L., Green-Church, K. B.,
and Krzycki, J. A. (2007). A natural genetic code expansion cassette enables transmissible
biosynthesis and genetic encoding of pyrrolysine. Proc. Natl. Acad. Sci. USA 104,
10211026.
Ludwig, M. L., and Matthews, R. G. (1997). Structure-based perspectives on B-12-depen-
dent enzymes. Annu. Rev. Biochem. 66, 269313.
Martin, B. D., and Finke, R. G. (1990). Co-C homolysis and bond dissociation energy
studies of biological alkylcobalamins: methylcobalamin, including a 10
15
Co-CH
3
homolysis rate enhancement at 25
o
C following one-electron reduction. J. Am. Chem.
Soc. 112, 24192420.
Matthews, R. G. (2001). Cobalamin-dependent methyltransferases. Acc. Chem. Res. 34,
681689.
Matthews, R. G., and Goulding, C. W. (1997). Enzyme-catalyzed methyl transfers to thiols:
the role of zinc. Curr. Opin. Chem. Biol. 1, 332339.
Maynard, E. L., and Lindahl, P. A. (1999). Evidence of a molecular tunnel connecting the
active sites for CO
2
reduction and acetyl-CoA synthesis in acetyl-CoA synthase from
Clostridium thermoaceticum. Journal of the American Chemical Society 121, 92219222.
Menon, S., and Ragsdale, S. W. (1998). Role of the [4Fe-4S] cluster in reductive activation
of the cobalt center of the corrinoid iron-sulfur protein from Clostridium thermoaceticum
during acetyl-CoA synthesis. Biochemistry 37, 56895698.
Menon, S., and Ragsdale, S. W. (1999). The role of an iron-sulfur cluster in an enzymatic
methylation reaction: methylation of CO dehydrogenase/acetyl-CoA synthase by the
methylated corrinoid iron-sulfur protein. J. Biol. Chem. 274, 1151311518.
Minot, G. R., and Murphy, W. P. (1926). Treatment of pernicious anaemia by a special diet.
J. Am. Med. Assn. 87, 470476.
Mitchell, H. K., Snell, E. E., and Williams, R. J. (1941). The concentration of folic acid.
J. Am. Chem. Soc. 63, 2284.
Myers, L. C., Cushing, T. D., Wagner, G., and Verdine, G. L. (1994). Metal-coordination
sphere in the methylated Ada protein-DNA co-complex. Chem. Biol. 1, 9197.
Naidu, D., and Ragsdale, S. W. (2001). Characterization of a three-component vanillate
O-demethylase from Moorella thermoacetica. J. Bacteriol 183, 32763281.
Norris, P. R., and Pratt, J. M. (1996). Methyl transfer reactions of protein-free Co
corrinoids. BioFactors 5, 240241.
Olah, G. A. (1993). Superelectrophiles. Angew. Chem. Int. Ed. 32, 767788.
Tetrahydrofolate and B
12
in Methyltransferases 321
Olteanu, H., and Banerjee, R. (2001). Human methionine synthase reductase, a soluble
P-450 reductase-like dual flavoprotein, is sufficient for NADPH-dependent methionine
synthase activation. J. Biol. Chem. 276, 3555835563.
Olteanu, H., Wolthers, K. R., Munro, A. W., Scrutton, N. S., and Banerjee, R. (2004).
Kinetic and thermodynamic characterization of the common polymorphic variants of
human methionine synthase reductase. Biochemistry 43, 19881997.
Peariso, K., Goulding, C. W., Huang, S., Matthews, R. G., and Penner-Hahn, J. E. (1998).
Characterization of the zinc binding site in methionine synthase enzymes of Escherichia
coli: The role of zinc in the methylation of homocysteine. J. Am. Chem. Soc. 120,
84108416.
Peariso, K., Zhou, Z. S., Smith, A. E., Matthews, R. G., and Penner-Hahn, J. E. (2001).
Characterization of the zinc sites in cobalamin-independent and cobalamin-dependent
methionine synthase using zinc and selenium X-ray absorption spectroscopy. Biochemistry
40, 987993.
Pejchal, R., and Ludwig, M. L. (2005). Cobalamin-independent methionine synthase
(MetE): a face-to-face double barrel that evolved by gene duplication. PLoS Biol. 3, e31.
Pfiffner, J. J., Calkins, D. G., ODell, B. L., Bloom, E. S., Brown, R. A., Campbell, C. J.,
and Bird, O. D. (1945). Isolation of an antianemia factor (vitamin Bc conjugate) in
crystalline form from yeast. Science 102, 228230.
Poston, J. M., Kuratomi, K., and Stadtman, E. R. (1964). Methyl-vitamin B
12
as a source of
methyl groups for the synthesis of acetate by cell-free extracts of Clostridium thermoaceti-
cum. Ann. N.Y. Acad. Sci. 112, 804806.
Pratt, J. M. (1999). The roles of Co, corrin, and protein: I. Co-ligand bonding and the trans
ligand effect. In Chemistry and biochemistry of B
12
, (R. Banerjee, ed.), Vol. 1,
pp. 73112. John Wiley and Sons, New York.
Pratt, J. M., Norris, P. R., Hamza, S. A., and Bolton, R. (1994). Methyl Transfer from
Nitrogen to Cobalt: Model for the B12-Dependent Methyl Transfer Enzymes. J. Chem.
Soc., Chem. Commun. 13331334.
Ragsdale, S. W. (2006). Metals and their scaffolds to promote difficult enzymatic reactions.
Chem. Rev. 106, 33173337.
Ragsdale, S. W., Lindahl, P. A., and Munck, E. (1987). Mossbauer, EPR, and optical studies
of the corrinoid/Fe-S protein involved in the synthesis of acetyl-CoA by Clostridium
thermoaceticum. J. Biol. Chem. 262, 1428914297.
Ragsdale, S. W., and Wood, H. G. (1985). Acetate biosynthesis by acetogenic bacteria:
Evidence that carbon monoxide dehydrogenase is the condensing enzyme that catalyzes
the final steps of the synthesis. J. Biol. Chem. 260, 39703977.
Ragsdale, S. W., Wood, H. G., and Antholine, W. E. (1985). Evidence that an iron-nickel-
carbon complex is formed by reaction of CO with the CO dehydrogenase from
Clostridium thermoaceticum. Proc. Natl. Acad. Sci. USA 82, 68116814.
Ram, M. S., and Riordan, C. G. (1995). Methyl transfer from a cobalt complex to Ni(tmc)
() yielding Ni(tmc)Me(): A model for methylcobalamin alkylation of CO dehydro-
genase. J. Am. Chem. Soc. 117, 23652366.
Ram, M. S., Riordan, C. G., Yap, G. P. A., Liable-Sands, L., Rheingold, A. L., Marchaj, A.,
and Norton, J. R. (1997). Kinetics and mechanism of alkyl transfer from organocobalt
(III) to nickel(I): Implications for the synthesis of acetyl coenzyme A by CO dehydroge-
nase. J. Am. Chem. Soc. 119, 16481655.
Rickes, E. L., Brink, N. G., Koniuszy, F. R., Wood, T. R., and Folkers, K. (1948).
Crystalline Vitamin B12. Science 107, 396397.
Riordan, C. (2004). Synthetic Chemistry and Chemical Precedents for Understanding the
Structure and Function of Acetyl Coenzyme ASynthase. J. Biol. Inorg. Chem. 9, 542549.
Rod, T. H., and Brooks, C. L., 3rd. (2003). How dihydrofolate reductase facilitates
protonation of dihydrofolate. J. Am. Chem. Soc. 125, 87188719.
322 Stephen W. Ragsdale
Sauer, K., and Thauer, R. K. (1997). Methanol:coenzyme M methyltransferase from
Methanosarcina barkeri-Zinc dependence and thermodynamics of the methanol:cob(I)
alamin methyltransferase reaction. Eur. J. Biochem. 249, 280285.
Schnyder, A., Darbre, T., and Keese, R. (1998). Methyl transfer from methanol to
Co-cobyrinate: a medel for the coenzyme B12 dependent methyltransferase. Angew.
Chem. Int. Ed. Engl. 37, 12831285.
Schrauzer, G. N., Deutsch, E., and Windgassen, R. J. (1968). The nucleophilicity of vitamin
B
12s
. J. Am. Chem. Soc. 90, 24412442.
Schrauzer, G. N., and Deutsch, E. J. (1969). Reactions of cobalt(I) supernucleophiles.The
alkylation of vitamin B12s, cobaloximes (I), and related compounds. J. Am. Chem. Soc.
91, 33413350.
Seravalli, J., Brown, K. L., and Ragsdale, S. W. (2001). Acetyl-Coenzyme A Synthesis From
Unnatural Methylated Corrinoids: Requirement for "Base-Off" Coordination at Cobalt.
J. Am. Chem. Soc. 123, 17861787.
Seravalli, J., and Ragsdale, S. W. (2000). Channeling of Carbon Monoxide during Anaero-
bic Carbon Dioxide Fixation. Biochemistry 39, 12741277.
Seravalli, J., Shoemaker, R. K., Sudbeck, M. J., and Ragsdale, S. W. (1999). Binding of (6R,
S)-methyltetrahydrofolate to methyltransferase from Clostridium thermoaceticum: Role of
protonation of methyltetrahydrofolate in the mechanismof methyl transfer. Biochemistry 38,
57365745.
Shibata, N., Masuda, J., Tobimatsu, T., Toraya, T., Suto, K., Morimoto, Y., and Yasuoka, N.
(1999). A new mode of B
12
binding and the direct participation of a potassium ion in
enzyme catalysis: X-ray structure of diol dehydratase. Structure 7, 9971008.
Sintchak, M. D., Arjara, G., Kellogg, B. A., Stubbe, J., and Drennan, C. L. (2002). The
crystal structure of class II ribonucleotide reductase reveals how an allosterically regulated
monomer mimics a dimer. Nat. Struct. Biol. 9, 293300.
Smith, A. E., and Matthews, R. G. (2000). Protonation state of methyltetrahydrofolate in a
binary complex with cobalamin-dependent methionine synthase. Biochemistry 39,
1388013890.
Smith, E. L. (1948). Purification of anti-pernicious anmia factors from liver. Nature 161,
638639.
Soares, J. A., Zhang, L., Pitsch, R. L., Kleinholz, N. M., Jones, R. B., Wolff, J. J., Amster, J.,
Green-Church, K. B., and Krzycki, J. A. (2005). The residue mass of L-pyrrolysine in
three distinct methylamine methyltransferases. J. Biol. Chem. 280, 3696236969.
Stich, T. A., Seravalli, J., Venkateshrao, S., Spiro, T. G., Ragsdale, S. W., and
Brunold, T. C. (2006). Spectroscopic Studies of the Corrinoid/Iron-Sulfur Protein
from Moorella thermoacetica. Journal of the American Chemical Society 128, 50105020.
Stokstad, E. L. R. (1943). Some properties of a growth factor for Lactobacillus casei. J. Biol.
Chem. 149, 573574.
Strickland, C. L., Windsor, W. T., Syto, R., Wang, L., Bond, R., Wu, Z., Schwartz, J.,
Le, H. V., Beese, L. S., and Weber, P. C. (1998). Crystal structure of farnesyl protein
transferase complexed with a CaaX peptide and farnesyl diphosphate analogue. Biochemistry
37, 1660116611.
Studer, A., Stupperich, E., Vuilleumier, S., and Leisinger, T. (2001). Chloromethane:
tetrahydrofolate methyl transfer by two proteins from Methylobacterium chlorometha-
nicum strain CM4. Eur. J. Biochem. 268, 29312938.
Stupperich, E., Eisinger, J. J., and Albracht, S. P. J. (1990). Evidence for a super-reduced
cobamide as the major corrinoid fraction in vivo and a histidine residue as a cobalt ligand
of the p-cresolyl cobamide in the acetogenic bacterium Sporomusa ovata. Eur. J. Biochem.
193, 105109.
Svetlitchnaia, T., Svetlitchnyi, V., Meyer, O., and Dobbek, H. (2006). Structural insights
into methyltransfer reactions of a corrinoid iron-sulfur protein involved in acetyl-CoA
synthesis. Proc. Natl. Acad. Sci. USA 103, 1433114336.
Tetrahydrofolate and B
12
in Methyltransferases 323
Svetlitchnyi, V., Dobbek, H., Meyer-Klaucke, W., Meins, T., Thiele, B., Romer, P.,
Huber, R., and Meyer, O. (2004). A functional Ni-Ni-[4Fe-4S] cluster in the mono-
meric acetyl-CoA synthase from carboxydothermus hydrogenoformans. Proc. Natl. Acad.
Sci. USA 101, 446451.
Tackett, S. L., Collat, J. W., and Abbott, J. C. (1963). The formation of hydridocobalamin
and its stability in aqueous solution. Biochemistry 2, 919923.
Tallant, T. C., Paul, L., and Krzycki, J. A. (2001). The MtsA subunit of the methylthiol:
coenzyme M methyltransferase of Methanosarcina barkeri catalyses both half-reactions of
corrinoid-dependent dimethylsulfide: coenzyme M methyl transfer. J. Biol. Chem. 276,
44854493.
Tan, X., Loke, H. K., Fitch, S., and Lindahl, P. A. (2005). The tunnel of acetyl-coenzyme a
synthase/carbon monoxide dehydrogenase regulates delivery of CO to the active site.
J. Am. Chem. Soc. 127, 58335839.
Tan, X., Sewell, C., Yang, Q., and Lindahl, P. A. (2003). Reduction and Methyl Transfer
Kinetics of the alpha Subunit from Acetyl Coenzyme A Synthase. J. Am. Chem. Soc. 125,
318319.
Thanbichler, M., Neuhierl, B., and Bock, A. (1999). S-methylmethionine metabolism in
Escherichia coli. J. Bacteriol 181, 662665.
Vetter, H., Jr, and Knappe, J. (1971). Flavodoxin and ferredoxin of Escherichia coli. Hoppe.
Seylers. Z. Physiol. Chem. 352, 433446.
Wassenaar, R. W., Keltjens, J. T., and van der Drift, C. (1998). Activation and reaction
kinetics of the dimethylamine/coenzyme M methyltransfer in Methanosarcina barkeri
strain Fusaro. Eur. J. Biochem. 258, 597602.
Wedemeyer-Exl, C., Darbre, T., and Keese, R. (1999). A model for the cobalamin-
dependent methionine synthase. Helvetica Chimica Acta. 82, 11731184.
Whipple, G. H., and Robscheit-Robbins, F. S. (1925). Blood regeneration in severe anemia.
II. Favorable influence of liver, heart and skeletal muscle in diet. Am. J. Physiol. 72,
408418.
Wilker, J. J., and Lippard, S. J. (1997). Alkyl transfer to metal thiolates: Kinetics, active
species identification, and relevance to the DNA methyl phosphotriester repair center of
Escherichia coli ada. Inorg. Chem. 36, 969978.
Wirt, M. D., Kumar, M., Ragsdale, S. W., and Chance, M. R. (1993). X-ray absorption
spectroscopy of the corrinoid/iron sulfur protein involved in acetyl-CoA synthesis by
Clostridium thermoaceticum. J. Am. Chem. Soc. 115, 21462150.
Wirt, M. D., Wu, J.-J., Scheuring, E. M., Kumar, M., Ragsdale, S. W., and Chance, M. R.
(1995). Structural and electronic factors in heterolytic cleavage: Formation of the Co(I)
intermediate in the corrinoid/iron-sulfur protein from Clostridium thermoaceticum.
Biochemistry 34, 52695273.
Yamanishi, M., Yamada, S., Muguruma, H., Murakami, Y., Tobimatsu, T., Ishida, A.,
Yamauchi, J., and Toraya, T. (1998). Evidence for axial coordination of 5,6-dimethyl-
benzimidazole to the cobalt atom of adenosylcobalamin bound to diol dehydratase.
Biochemistry 37, 47994803.
Zhao, S., Roberts, D. L., and Ragsdale, S. W. (1995). Mechanistic studies of the methyl-
transferase from clostridium thermoaceticum: origin of the pH dependence of the methyl
group transfer from methyltetrahydrofolate to the corrinoid/iron-sulfur protein. Biochem-
istry 34, 1507515083.
Zheng, D., Darbre, T., and Keese, R. (1999). Methanol and dimethylaniline as methylating
agents of Co(I) corrinoids under acidic conditions. J. Inorg. Biochem. 73, 273275.
Zhou, Z. H. S., Peariso, K., PennerHahn, J. E., and Matthews, R. G. (1999). Identification
of the zinc ligands in cobalamin- independent methionine synthase (MetE) from Escher-
ichia coli. Biochemistry Usa. 38, 1591515926.
324 Stephen W. Ragsdale
C H A P T E R E L E V E N
Methyltetrahydrofolate in
Folate-Binding Protein Glycine
N-Methyltransferase
Zigmund Luka*
Contents
I. Introduction 326
II. 5-Methyltetrahydrofolate in Folate Metabolism 326
A. Folate uptake, metabolism, and catabolism 326
B. 5-Methyltetrahydrofolate, methyl trap 328
III. Folate-Binding Enzymes 330
IV. Glycine N-Methyltransferase 331
A. GNMT genes and proteins 331
B. Enzyme kinetics and activity regulation 333
V. GNMT as Methyltetrahydrofolate-Binding Protein 335
A. Inhibition of GNMT by folate 336
B. Role of GNMT in folate and one-carbon metabolism 336
C. GNMT mutations in human patients and
GNMT knockout mouse model 337
D. Crystal structure of GNMTfolate complex
and mechanism of inhibition 339
VI. Conclusions 340
Acknowledgments 341
References 341
Abstract
In mammals, folate is used as a carrier of one-carbon units (C
1
) in nucleic acids
metabolism and biological methylation. Among all forms of folate the most
abundant is 5-methyltetrahydrofolate (5-CH
3
-THF), which is of exceptional
importance. Its distinctive role among other forms of folate is in its dual
function. As a C
1
carrier it is used for synthesis of methionine by remethylation
of homocysteine. In addition, 5-CH
3
-THF is bound to and inhibits glycine-N-
methyltransferase (GNMT). GNMT is one of the key enzymes in methionine and
S-adenosylmethionine (AdoMet) metabolism. It removes excess AdoMet by
Vitamins and Hormones, Volume 79
#
2008 Elsevier Inc.
ISSN 0083-6729, DOI: 10.1016/S0083-6729(08)00411-1 All rights reserved.
* Department of Biochemistry, Vanderbilt University School of Medicine, Nashville, Tennessee 37232
325
using it for methylation of glycine. The interaction of 5-CH
3
-THF and GNMT was
proposed as an important regulatory mechanism in AdoMet metabolism and
biological methylation. The recent discovery of human individuals with mutant
GNMT and the study of a mouse model with the GNMT gene knocked out
showed that inactivation of that enzyme, indeed, has a significant impact on
AdoMet levels in the liver and plasma. The crystal structure of GNMTcomplexed
with 5-CH
3
-THF revealed that there are two folate molecules bound to one
tetrameric form of GNMT, which is a basis for establishing of mechanism of
inhibition of GNMT. The role of GNMT as a folate-binding protein and how it
affects one-carbon folate metabolism is discussed. 2008 Elsevier Inc.
I. Introduction
Living organisms need folate as enzyme cofactors to carry C
1
from a
number of sources for biosynthesis of nucleic acids and methylation. After
absorption from the diet in the intestine, transferring to circulation and
transport into the cells, the folates are bound by folate-binding enzymes so
there is almost no free folate in the cells (Wagner, 1995). In the liver, the most
abundant form of folate is methyltetrahydrofolate and changes in its amount
are associated with several metabolic conditions. Only one enzyme, methio-
nine synthase (MS), uses methyltetrahydrofolate as a substrate but this is not a
major methyltetrahydrofolate-binding protein. In the liver and several other
tissues, the major methyltetrahydrofolate-binding protein is the enzyme,
glycine N-methyltransferase (GNMT). This enzyme, in addition to its
important biological role of removing excess S-adenosylmethionine by
methylation of glycine, also binds methyltetrahydrofolate without using it
for any enzymatic reaction. Moreover, methyltetrahydrofolate is a potent
inhibitor of GNMT. The high level of GNMT in the liver controls the
availability of the folate for other enzymes and use of AdoMet for methylation
reactions since all forms of folate are interconvertible. For these reasons more
attention should be paid to the biological role of this particular
methyltetrahydrofolate-binding protein/enzyme. In this chapter, the latest
data on the role of GNMT as folate-binding regulatory protein are presented.
II. 5-Methyltetrahydrofolate in Folate
Metabolism
A. Folate uptake, metabolism, and catabolism
A comprehensive review on folate metabolism with discussion of roles of
different folate forms is published in this issue (Chapter 1 by Stover). In this
chapter, the biological role of one form of folate, methyltetrahydrofolate,
will be the main subject, although the more important steps in folate
metabolism will be briefly discussed.
326 Zigmund Luka
Mammals get folate from the diet as a mixture of different forms (Cook,
2001; McKillop et al., 2006; Seyoum and Selhub, 1998). The major form of
folate in the diet is 5-CH
3
-THF, which is, for example, 85% in egg yolk,
43% in spinach, 46% in yeast (McKillop et al., 2006), 36% in cows liver, and
58% in orange juice (Seyoum and Selhub, 1998). The structure of 5-CH
3
-
THF is shown in Fig. 11.1. Other folate forms differ in substitutions at C5
and C10 carbon atoms on the pterin and in the number of glutamate
residues linked to p-aminobenzoate (PABA). The number of glutamate
residues varies from 1 to 8 depending on the food (Cook, 2001; McKillop
et al., 2006; Seyoum and Selhub, 1998).
Folate uptake is a complex process, which is discussed in published
reviews (Halsted, 1979; Matherly and Goldman, 2003; Pacha, 2000; Said,
2004; Said and Mohammed, 2006; Sirotnak and Tolner, 1999). There are
Figure 11.1 5-Methyltetrahydrofolate(5-CH
3
-THF) and scheme of C
1
unit transfer.
The structure of 5-CH
3
-THF is shown in the upper part. The scheme at the bottom
shows the sources of C
1
units for THF loading and the major folate forms used for
synthesis of methionine, deoxythymidine 5
0
-phosphate (dTMP), and purines.
GNMT-methyltetrahydrofolate 327
two main steps of folate transport. The first step begins with folate absorption
by the small intestine epithelium and transfer of folate into mucosa cells.
In the brash border of mucosal cells, all polyglutamate forms are converted
to monoglutamates by action of a specific enzyme, folylpoly-g-glutamate
carboxypeptidase (Shane, 1995). The monoglutamate forms of folates are
transported through intestine epithelium cells into the circulation by a passive
mechanism at high concentration of folate and by the reduced folate carrier
(RFC) at lower folate levels. In the second step folate monoglutamates are
transported from circulation into the cells by the RFCand by folate receptors.
Folate is used in C
1
metabolism in all tissues but most folate is metabolized
in the liver (Clifford et al., 1990; Cook, 2001). In the liver cells the folate pool
is almost equally divided between cytosol and mitochondria. In both compart-
ments the first step of metabolismis the addition of glutamate residues to folate
monoglutamates by folylpoly-g-glutamate synthase (Shane, 1995). In the rat
the predominant form is pentaglutamate (Shin et al., 1972), while in human it
is hexa- and heptaglutamates (Foo et al., 1982).
Folates in the cell serve as carriers of C
1
groups. In the very simplified
scheme in Fig. 11.1 the main sources of C
1
groups and their use in the cell is
presented. The C
1
groups are added to tetrahydrofolate (THF) from serine,
formate, choline, glycine, sarcosine, andfromhistidine metabolites bydifferent
cytosolic and mitochondrial enzymes.
Loading of THF with different C
1
groups results in synthesis of different
forms of folate, which are using for synthesis of purines, deoxythymidine
5
0
-phosphate (dTMP), and methylation as shown in Fig. 11.1. In the cell the
distribution of the folate forms is maintained relatively constant. In the rat
liver cytosolic folate pool there are four major forms: THF, 27% of total
folate; 5-CH
3
-THF, 45%; 10-formylTHF, 19.3%; 5-formylTHF, 8.7%,
and small amounts of other forms: 5-formiminoTHF, 5,10-methenylTHF,
5,10-methyleneTHF, and dihydrofolate (Cook, 2001). Detailed folate
metabolism is discussed in another chapter of this issue (Chapter 1 by
Stover)) and in other reviews (Appling, 1991; Cook, 2001; Wagner, 1995).
Catabolism of folates includes three main pathways. Free folate monoglu-
tamates are passively excreted from the cells into the circulation. Free folate
polyglutamates are hydrolyzed to monoglutamates by action of g-glutamyl
hydrolase (GGH) and subsequently excreted as monoglutamates. Free folates
also undergo oxidative degradation releasing different products depending on
the degradation scheme (Suh et al., 2001).
B. 5-Methyltetrahydrofolate, methyl trap
All forms of folates are used in specific enzymatic reactions and the impor-
tance of most of them for cell metabolism is well studied. Among them one
form is of exceptional importance. This is 5-CH
3
-THF, the major form of
folate in cytosol. The 5-CH
3
-THF is synthesized from 5,10-methylene-THF
328 Zigmund Luka
by methylene-THF reductase (MTHFR). Its exceptional role is in its dual
function. First, it is used for synthesis of methionine by remethylation of
homocysteine by MS thus maintaining a constant supply of methionine for
protein synthesis and methylation reactions (Fig. 11.2).
However, if utilization of 5-CH
3
-THF by MS is somehow (i.e., by B
12
deficiency) interrupted more and more folate accumulates as 5-CH
3
-THF.
This happens because this formof folate could be converted into THF only by
MS and 5,10-methylTHF reductase is irreversible. This is called the methyl
trap theory (Herbert and Zalusky, 1962). Accumulation of 5-CH
3
-
THF results in severe abnormalities in folate metabolism and human disease
(Shane and Stokstad, 1985).
Methionine
AdoMet
1
2
3
4
5
6
AdoHcy
Adenosine
Homocysteine
Cystathionine
Cysteine and Pyruvate
THF
Diet
Folate pool
Glycine
5-CH
3
-THF
5, 10-CH
2
-THF
5-CH
3
-THF
inhibition
AdoMet
inhibition
Methylation
reactions
Sarcosine
Figure 11.2 Participation of glycine-N-methyltransferase (GNMT) in methionine and
folate metabolism. Numbers at arrows show enzymes involved in synthesis and regula-
tion of methionine metabolism: 1-methionine adenosyltransferase; 2-a variety of
methyltransferases; 3-GNMT; 4-S-adenosylhomocysteine hydrolase; 5-methionine
synthase; 6- methylene tetrahydrofolate reductase. Inhibition of reaction 6 by AdoMet
and reaction 3 by 5-CH
3
-THF is shown.
GNMT-methyltetrahydrofolate 329
The importance of 5-CH
3
-THF for overall metabolism was studied
experimentally by using the gene knockout technique. Two mouse models
were developed: one with MS knocked out (Swanson et al., 2001) and
another with MTHFR knocked out (Chen et al., 2001). Both models
showed severe effects of changed of 5-CH
3
-THF levels. The MS knockout
appeared to be lethal and no homozygous MS-deficient mice were
obtained. The detailed study of heterozygous animals led to the conclusion
that the reason for lethality was a methyl trap. The MTHFR-deficient mice
exhibited severe abnormalities in early development. In cases of those
animals that survived, the level of 5-CH
3
-THF in the liver was more than
20-fold lower, the level of AdoMet slightly lower and the level of AdoHcy
in the liver was 410 times higher with a number of physiological abnorm-
alities. Both studies clearly showed that the level of 5-CH
3
-THF must be
carefully maintained in the cell for methionine synthesis.
The second role of 5-CH
3
-THF was unexpected when it was first
discovered: it serves as an allosteric regulator of GNMT. The enzyme
does not use folate as a substrate for any enzymatic reaction but it is one
of the most important regulatory proteins in methionine and S-adenosyl-
methionine metabolism. Thus, 5-CH
3
-THF is used not only as a substrate
for homocysteine remethylation reaction but also as a regulatory element in
cellular methylation.
III. Folate-Binding Enzymes
Once folate is transferred into the cell all folate forms are bound to
folate-metabolizing enzymes. The level of folate-binding proteins in the cell
is extremely high compared to other proteins expressed. The major folate-
binding enzymes (both cytosol and mitochondria) are: 10-formylTHF
dehydrogenase (1.2% of pig liver total protein, Min et al., 1988), serine
hydroxymethyltransferase (0.8% in rabbit liver, Schirch and Gross, 1968),
dimethylglycine dehydrogenase (0.8% in rat liver, Wittwer and Wagner,
1980), C
1
-THF synthase (0.5% in sheep liver, Paukert et al., 1976), and
GNMT (13% in rat and rabbit livers, Heady and Kerr, 1973; Ogawa and
Fujioka, 1982a). As a result, in the cell there is almost no free folate because
there are more folate-binding sites on the proteins then there are folate
molecules (Cook, 2001; Henderson, 1990; Kim et al., 1996; Wagner, 1982,
1995). This obviously depends on composition of the pool of the folate
forms. THF is used for C
1
-group loading by a number of enzymes and
therefore must be available in large quantity. For DNA synthesis,
10-formylTHF and 5,10-methyleneTHF must be available and their level
in the cell is also high.
330 Zigmund Luka
IV. Glycine N-Methyltransferase
The GNMT activity was detected first by Blumenstein and Williams
(1960) in their search for a way for direct transfer of the methyl group to
glycine from S-adenosylmethionine as a methyl donor. Such an activity was
detected in all liver extracts they tested by formation of radioactively labeled
sarcosine from radioactive glycine. The authors did not proceed with enzyme
purification, however. An enzyme catalyzing methylation of glycine was
identified and purified from rabbit liver by Heady and Kerr (1973) and it
was named as glycine N-methyltransferase and abbreviated as GNMT.
A. GNMT genes and proteins
For a long time GNMT activity was found only in mammals, although
genes with homology to mammalian GNMT are present in many species
genomes, for example, Zebrafish or red flour beetle. Glycine methylation in
other organisms was reported only for cyanobacteria Aphanothece halophytic
(Waditee et al., 2003) by a monomeric enzyme with low homology to rat
GNMT. Since it is not known whether these other forms of GNMT bind
folate in this review only mammalian GNMTs are considered. The GNMT
proteins are encoded by similar genes in all species, from which they were
cloned and sequenced: rat (Ogawa et al., 1987), mouse (Luka and Wagner,
GB accession number AF 325352, 2001), human (Chen et al., 2000), and
chimpanzee (GB accession number, AACZ02071362). All GNMT genes are
relatively small, about 5 kb with a coding sequence of 6 exons of 100200
nucleotides (Fig. 11.3).
The best-studied gene is that from human because of its potential as a
cancer gene marker (Chen et al., 2000; Tseng et al., 2003). It was found that
the transcriptional site is positioned at 800 nucleotide (Chen et al., 2000).
This was unusual because in rat, the transcriptional site was located just
upstream of an ATG codon in the first exon (Ogawa et al., 1987). There
were several polymorphic sites found in human GNMT gene (Tseng et al.,
2003). In the 5
0
upstream sequence the most common are a C/T at 45
nucleotide, DTTACA deletion/insertion at 991, and GA repeats at 1259
nucleotides. Regarding these polymorphisms the human GNMT genes
consists of two main alleles: one with 11 GA repeats and another with
16 GA repeats. One allele (1289 T-allele) carries 11 GA repeats, a DTTACA
insertion and thymidine at 45, while another allele (1289 C-allele) carries
16 GA repeats, a sequence DTTACA is deleted and cytidine at 45 position.
Regulation of GNMT gene expression has not been studied except for
a few experiments (Tseng et al., 2003). From those experiments with a
gene reporter system it was found that the 11 GA allele (1289 T-allele)
is expressed about 1.6 times more efficiently than 16 GA allele when
GNMT-methyltetrahydrofolate 331
5
0
-upstream sequence to 2034 nucleotides were used in a gene reporter
vector. What was found for certainty was that the level of mRNA for
GNMT was higher in liver, kidney, pancreas, and prostate (Chen et al.,
2000), that is, similar to results obtained by analysis of GNMT protein
abundance (see below).
GNMTenzyme was first purified fromrabbit liver and immediately it was
found that that enzyme was unusually highly abundant in that organ with an
estimated level of up to 3%of total soluble proteins in the cytosol (Heady and
Kerr, 1973). In the same study, GNMTactivity was found also in kidney and
pancreas. No GNMTactivity was detected in fetal liver and in three kinds of
cancer cells: Novikoff hepatoma, Morris hepatoma, and Ehrlich ascites cells
(Heady and Kerr, 1973, 1975). Significantly, lower expression of GNMT
was found in hepatocellular carcinoma cells in the later studies (Liang et al.,
2005; Liu et al., 2003). Similar to the rabbit, the highest level of this enzyme in
the rat was found in liver (Ogawa and Fujioka, 1982a), where it consists of
about 1% of soluble proteins, and lower but still a significant level in rat
pancreas (Yeo and Wagner, 1992, 1994).
After initial studies on GNMT properties were done on rabbit liver
protein, more detailed analysis was performed on the rat liver enzyme.
GNMT from that source was purified to homogeneity (Ogawa and Fujioka,
1982a) and protein properties were studied using protein chemistry
such as analysis of reactivity of sulfhydryl groups (Fujioka et al., 1987) and
probing of protein conformation by arginine modification (Konishi
and Fujioka, 1987). Protein sequence was determined by cloning and
Figure 11.3 Glycine-N-methyltransferase (GNMT) gene and protein. At the top the
scheme of the mouse GNMT gene (GeneBank accession number AF 325352) is shown.
Numbers from 1 to 6345 indicate the numbers of nucleotides. Below are aligned amino
acid sequences of mouse and human GNMT (GeneBank codes D89664 and
NM018960). Different amino acid residues are in boxes.
332 Zigmund Luka
sequencing of GNMT cDNA (Ogawa et al., 1984, 1987). Native and recom-
binant GNMTs were compared by kinetics, limited proteolysis and AdoMet
binding (Konishi and Fujioka, 1988). Ogawas research group reported the
first crystal structure of GNMT, which will be discussed below. GNMT from
pancreas was also purified to homogeneity and its kinetic properties were
studied (Yeo and Wagner, 1992). Recombinant human and mouse GNMTs
were expressed in Escherichia coli and their kinetic properties, stability, and
crystal structures were studied (Luka and Wagner, 2003a; Luka and Wagner,
2003b; Pakhomova et al., 2004).
It was found that GNMT is a tetrameric protein consisting of four
identical subunits of about 32.5 kDa, which depends on the protein origin.
The cDNAs for GNMT from human, rabbit, pig, and mouse were cloned
and sequenced (Aida et al., 1997; Chen et al., 1998; Ogawa et al., 1993) from
which amino acid sequences were derived. It appeared that GNMTs are
very similar with amino acid sequence similarity of about 90%. As an
example, in Fig. 11.3 sequences of GNMTs from mouse and human are
aligned. All sequenced proteins contain 292294 amino acid residues with
no signs of any unusual features. The theoretical isoelectric points are about
7.17.5, but experimentally, the pI value for native rat enzyme was found to
be 6.4 (Ogawa and Fujioka, 1982a).
Rat liver GNMT is phosphorylated in vivo and could be phosphorylated
in vitro by cAMP-dependent protein kinase (Wagner et al., 1989). This was
confirmed by hepatocyte studies (Moller et al., 2003). The studies on rat liver
and recombinant GNMTs showed that there are several phosphorylated
serine residues, but the total level of phosphorylation is very low (Luka
et al., 2006a). It was found that the only residues undergoing phosphorylation
are serines and in rat GNMT those serine residues are Ser9, Ser71, Ser139,
Ser182, and Ser241. The Ser9 was phosphorylated in vitro when cAMP-
dependent protein kinase was used (Luka et al., 2006a; Wagner et al., 1989).
The GNMT crystal structure was solved for rat, mouse, and human
recombinant proteins either as an apoprotein on in complex with AdoHcy,
AdoMet, AdoMet and acetate, or as a complex with 5-CH
3
-THF (Fu et al.,
1996; Huang et al., 2000; Luka et al., 2007; Pakhomova et al., 2004;
Pattanayek et al., 1998; Takata et al., 2003). It appeared that in the crystal,
GNMT is organized as flat-shaped tetramer with numerous interactions
between subunits. Each subunit consists of an active center for reaction of
AdoMet and glycine inside of the globular part.
B. Enzyme kinetics and activity regulation
Kinetics of GNMT was studied on native liver and pancreatic enzymes and
on recombinant enzyme from different sources. Some kinetic properties of
native GNMT were obtained by using crude liver extracts from pig and
human (Ogawa et al., 1993). Measuring of the dependence of reaction
GNMT-methyltetrahydrofolate 333
velocity on concentration of AdoMet and glycine revealed that K
m
for
AdoMet is in the range of 0.030.2 mM and for glycine in the range of
220 mM(Luka and Wagner, 2003a; Ogawa and Fujioka, 1982a; Pakhomova
et al., 2004). The k
cat
values were found in the range of 3596 min
1
(Luka
and Wagner, 2003a; Ogawa and Fujioka, 1982a; Pakhomova et al., 2004).
GNMT is the only methyltransferase, which is a tetrameric protein. Based on
that fact one could assume that there should be some kind of cooperativity
in the response tothe substrate concentration. Such cooperativity was found in
rat native enzyme toward AdoMet but not toward glycine: in the latter case
dependence of velocity on glycine concentration is purely hyperbolic (Ogawa
and Fujioka, 1982a). Interestingly enough, that AdoMet-cooperativity was
found in case of native liver enzymes only (Konishi and Fujioka, 1988; Ogawa
et al., 1993): recombinant rat, mouse, and human enzymes showed no coop-
erativity toward AdoMet (Konishi and Fujioka, 1988; Luka and Wagner,
2003a; Pakhomova et al., 2004). That finding was explained as the result of
acetylation of N-terminal valine in native protein and absence of acetyl group
on valine in recombinant protein (Ogawa et al., 1997). The latter conclusion
was drawn on the basis of removing of the N-terminal 7 amino acids, not just
deacetylation of the native protein or acetylation of recombinant enzyme.
In another case, rabbit liver enzyme (Kloor et al., 2004) showed no AdoMet-
cooperativity with the rat liver enzyme.
Analysis of GNMT crystal structure allowed proposing the mechanism of
enzymatic reaction. This mechanism is discussed in detail in Takata et al.
(2003). The reaction begins with AdoMet binding to the active center. To
make this possible, the N-terminal loop of 120 amino acids in apoprotein
should be removed from its position blocking the entrance to active center.
The N-terminal fragment is exposed into solvent as a result of major confor-
mation rearrangements. After transferring the methyl group from the posi-
tively charged AdoMet to glycine via an S
N
2 reaction, the neutral sarcosine
and AdoHcy do not interact strongly with protein moiety and are released.
A few residues play critical roles in the enzymatic reaction: Tyr21, Tyr33,
Gly137, Asn138, His142, Arg175, and Tyr194 (residue numbers for rat
GNMT). The mechanism of the transfer of the methyl group from AdoMet
to glycine by GNMTwas evaluated by theoretical studies (Soriano et al., 2006;
Velichkova and Himo, 2005) and it was found that generally it is in good
agreement with that of proposed by Takata et al. withthe exceptionthat Tyr21
most likely is not involved in reaction (Velichkova and Himo, 2005).
In the very first study of GNMT from rabbit liver, it was found that one
product of the reaction AdoHcy is a moderate inhibitor of reaction with
constant of inhibition (K
i
) of 40 mM (Heady and Kerr, 1973). The other
product, sarcosine, is not an inhibitor.
The enzyme activity of GNMT is changed in response to a number of
factors. There are two levels of GNMT activity regulation: one is genetically
determined and results in tissue specificity of GNMT expression. Another
334 Zigmund Luka
level of GNMT activity regulation is a result of changes of activity and gene
expression in response to diet, hormone status, drug administration, and so
on. When rats were fed a high-methionine diet (3% methionine), it resulted
in an increase in GNMT activity and abundance of GNMT protein in the
liver (Ogawa and Fujioka, 1982b). This result was confirmed by Rowling
et al. (2002). GNMT activity is changed in response to some other treatments,
like retinoids (McMullen et al., 2002) or ethanol administration (Villanueva
and Halsted, 2004).
The hormonal status is an important factor in methyl group metabolism
as recently summarized (Williams and Schalinske, 2007). That conclusion
was based also on studies of changes of GNMT activity with hormonal
status and treatment. In an early study, it was found that activity of GNMT
in rat and mice liver and kidney increased with aging (Mays et al., 1973).
The increase was of about 50% of initial activity at 312 months in rat and
almost 100% increase in mice between 12 and 30 months. Later it was found
that GNMT expression is regulated by growth hormone (Aida et al., 1997;
Brown-Borg et al., 2005; Uthus and Brown-Borg, 2003).
Other hormones, glucagon and glucocorticoids, affect GNMT activity
(Rowling and Schalinske, 2003). Glucagon treatment of rat increased
GNMT activity by 25%, which was explained as a result of change of
metabolites in the transsulfuration pathway ( Jacobs et al., 2001). In adrenal-
ectomized rats administration of dexamethazone increased GNMT activity
3.6-fold (Rowling and Schalinske, 2003). In alloxan-induced diabetic sheep
the level of GNMT transcript increased 65-fold (Xu and Snoswell, 1985).
That level of increase might be an exception since similar experiment with
rats showed only about twofold increase in GNMT activity in the liver (Yeo
and Wagner, 1994). In diabetic rats both a folate-enriched diet and folate
deficiency resulted in increased GNMT activity (Nieman et al., 2006).
Phosphorylation is another possible way to regulate GNMT activity. By
phosphorylation of GNMT with protein kinase A, activity of GNMT
significantly increased (Wagner et al., 1989). The question still remains
whether GNMT phosphorylation in vivo results in such changes of confor-
mation or some target protein binding, which could be of importance in
terms of the biological role of GNMT.
Among all factors affecting GNMT activity the most important appeared
to be inhibition of GNMT by folate (Wagner et al., 1985; Yeo et al., 1999),
which will be discussed in detail below.
V. GNMT as Methyltetrahydrofolate-
Binding Protein
The fact that GNMT is somehowparticipating in regulation of the ratio
of AdoMet/AdoHcy and level of methylation in the cells was implied since
discovery of GNMT (Heady and Kerr, 1973); however, the mechanism of
GNMT-methyltetrahydrofolate 335
that regulation was not known. Establishment of such a mechanism began
with the discovery of GNMT as major rat liver cytosolic folate-binding
protein. In 1977 Zamierowski and Wagner reported that injected radioactive
folic acid into rat mostly bound to one of several unidentified rat liver
cytosolic proteins.
That protein was purified to homogeneity and found that it binds almost
exclusively 5-CH
3
-THF. Antibodies to that protein were produced (Suzuki
and Wagner, 1980) and radioimmunoassay showed that that this new folate-
binding protein was most abundant in rat liver cytosol. It was also expressed
in kidney, prostate, and some other organs (Cook and Wagner, 1981).
A surprising discovery was made when it was found that the unknown
folate-binding protein was, in fact, GNMT (Cook and Wagner, 1984).
A. Inhibition of GNMT by folate
The detailed studies on folate binding by GNMT showed that folates inhibit
GNMT activity (Wagner et al., 1985). The most potent inhibitor is 5-CH
3
-
THF. Later, inhibition was studied more precisely on rat liver and pancre-
atic GNMT (Yeo et al., 1999). It was found that inhibition of GNMT
activity by 5-CH
3
-THF-Glu
5
is noncompetitive regarding substrates, Ado-
Met and glycine. Inhibition of liver and pancreatic enzymes is similar. For
rat liver enzyme, 50% activity inhibition was observed at 2 mM folate
concentration at pH 7.5 and at 1 mM at pH 9.0. The binding constants
were determined and it was found that K
d
depends on pH: at pH 7.4 the K
d
value was determined as 0.095 mM, while at pH 9.0 that value was much
higher of 1.9 mM. Interestingly, the cooperativity toward AdoMet was
found much stronger at pH 7.4 than at pH 9.0 with correspondent
Hill coefficients of 2.0 and 1.5 at 5-CH
3
-THF concentration of 0.2 mM.
The cooperativity is increased with folate concentration.
B. Role of GNMT in folate and one-carbon metabolism
Discovery of folate binding by GNMT and inhibition of GNMT activity by
folates raised the question of what is the biological importance of such an
interaction and inhibition. In the working hypothesis of Wagner and his
colleagues (Balaghi et al., 1993; Wagner et al., 1985) proposed the relation-
ship between methyl group and folate one-carbon pool metabolisms with a
crucial regulatory role of GNMT.
In the simplified scheme in Fig. 11.2 the essential metabolic pathways and
enzymatic reactions, which are involved in methionine and folate metabolism
regulation by GNMTare presented. Methionine comes fromthe diet and it is
used for synthesis of proteins and AdoMet by methionine adenosine transfer-
ase (MAT, reaction 1). AdoMet is used by about 5060 methyltransferases in
human (Clarke and Banfild, 2001). In Fig. 11.2, all of the methyltransferases
336 Zigmund Luka
are shown as reaction 2, but for GNMT it is shown as reaction 3. One of the
products of methylation reactions is AdoHcy which undergoes further
metabolismvia transsulfuration reactions. AdoHcy is hydrolyzed to adenosine
and homocysteine by S-adenosylhomocysteine hydrolase (reaction 4).
Homocysteine can be converted back to methionine by transferring a methyl
group from 5-CH
3
-THF by MS (reaction 5). In turn, 5-CH
3
-THF is
synthesized from 5,10-methylene-THF by MTHFR (reaction 6) and that
reaction is inhibited by AdoMet.
The levels of methionine and AdoMet are determined by diet and by the
availability of methyl groups from 5-CH
3
-THF, which may vary signifi-
cantly. Since the level of AdoMet can affect biological methylation, it must
be thoroughly regulated. The proposed regulation mechanism explains how
it is achieved, at least in the liver, where most of the AdoMet and folate are
metabolized. In the first scenario, when methionine level in the diet and in
the cell is higher than normal, the level of AdoMet is also abnormally high
because there is no feedback between level of AdoMet and MAT activity.
A high level of AdoMet results in increase in inhibition of MTHFR (reaction
6) and less 5-CH
3
-THF is synthesized. This, in turn, leads to activation of
GNMT because of less inhibition by 5-CH
3
-THF (reaction 3). That results
in increased use of AdoMet for glycine methylation and funneling of excess
AdoMet through the transsulfuration pathway (homocysteine-cystathionine-
cysteine and pyruvate) shown in Fig. 11.3.
When the methionine level becomes lower than normal the level of
AdoMet is also decreased below normal and this must be prevented. Again,
this is achieved by activation/inhibition of MTHFR and GNMT. The
MTHFR is activated because concentration of its inhibitor, AdoMet,
decreased. Therefore, more 5-CH
3
-THF is synthesized, more homocyste-
ine is remethylated to methionine, and more AdoMet synthesized. At the
same time increased concentration of 5-CH
3
-THF inhibits GNMT, less
AdoMet is used for glycine methylation and becomes available for other
methylation reactions (reaction 2).
C. GNMT mutations in human patients and
GNMT knockout mouse model
The proposed mechanism of AdoMet level regulation via GNMT was
confirmed in vivo by: (a) search for humans with genetically inactivated
GNMT and (b) by using a mouse model with gene for GNMT knocked
out. If the proposed scheme of regulation is true then absence or low
activity of GNMT should result in high level of AdoMet and methionine
with normal or lower levels of AdoHcy.
Exploring the first approach three patients were found with extremely
high plasma levels of methionine and AdoMet and normal levels of AdoHcy
(Augoustides-Savvopoulou et al., 2003; Luka et al., 2002; Mudd et al., 2001).
GNMT-methyltetrahydrofolate 337
Two of these patients were children from one Italian family and a third
patient was a child from Greece. All of them have very high level of AdoMet
in the plasma (2.22.7 mM with normal level about 100 nM) and methionine
(4001000 mM with normal level 1545 mM), but normal level of AdoHcy.
All patients were characterized with mild liver disease.
GNMT genes of all three patients were sequenced. In the case of the
Italian family DNA from parents and both children was also analyzed.
Indeed, as was expected, GNMT genes from all patients carried missense
mutations in the exons. The brother and sister from the Italian family carry
the two mutations: one (C/T in CTT codon for Leu49) in exon 1 and
another (C/A in codon CAT for His176) in exon 4. The GNMT gene of
the third patient carries an A/G mutation in codon AAT for Asn140. These
mutations cause change of Leu49 to proline, Asn140 to serine, and His176
to asparagines, respectively.
Activity of all mutant GNMTs were assayed in vitro after cDNAs with
corresponding mutations were cloned and expressed in an E. coli expression
vector (Augoustides-Savvopoulou et al., 2003; Luka and Wagner, 2003c).
It was shown that all mutations inactivate GNMT to different extents: the
N140S mutant possess only traces of WT GNMT activity, L49P is inacti-
vated by 90%, but the H176N mutant activity decreased only 25% regard-
ing V
max
(Luka and Wagner, 2003c). The K
m
values for AdoMet and
glycine also were changed compared to WT GNMT. These data are in
excellent agreement with proposed scheme of regulation of the levels of
methionine and AdoMet by GNMT.
The second strategy to show the role of GNMT in vivo also was explored
by developing the mouse model with GNMT knocked out by using
gene targeting (Luka et al., 2006b). In that mouse model exon 1 and the 5
0
-
regulatory sequence of GNMT gene, which carries at least part of putative
promoter in targeted allele, was replaced via homologous recombination with
NEO (neomycin phosphotransferase) gene.
Transgenic mice showed no difference in appearance and growth rate at
least for 36 months. GNMT activity assay and Western blotting showed
that transgenic animals, indeed, possessed no GNMT activity and no
GNMT protein. The levels of main metabolites were exactly as was
expected for inactivated GNMT and as was found in human patients: the
concentration of methionine increased from 100 nmol/g liver in WT
animals to about 700 nmol/g, the concentration of AdoMet increased
from 37 nmol/g in WT animals to about 1334 nmol/g in knockout
mice, while the level of AdoHcy slightly decreased from 1215 nmol/g to
35 nmol/g liver (Luka et al., 2006b). These changes in concentrations of
AdoMet and AdoHcy resulted in an increase in the AdoMet/AdoHcy ratio
from about 3300, that is, 100-fold. These changes mean that concentration
of the substrate for almost all methylases, AdoMet, increased enormously.
At the same time, the concentration of AdoHcy, which is an inhibitor for
338 Zigmund Luka
most of the methyltransferases that use AdoMet, decreased. That should
result in significant increase in methylation reactions in the cells.
D. Crystal structure of GNMTfolate complex
and mechanism of inhibition
Biochemical and genetic studies showed the regulatory role of GNMT in
methionine and folate metabolism. The mechanism of GNMT inhibition
by folate was proposed by solving crystal structure of the GNMTfolate
complex by Luka et al. (2007). The rat recombinant GNMT and 5-CH
3
-
THF (monoglutamate) were used in their study. In binding experiments, it
was found that there were two 5-CH
3
-THF molecule bound to one
molecule of tetrameric form of GNMT in solution at pH 7.5. That was
different from earlier experiments it was found that at pH 7.4 only one
molecule of 5-CH
3
-THF-Glu
5
(pentaglutamate) was bound to mole of
tetrameric GNMT (Yeo et al., 1999). The reason for that discrepancy was
explained as a result of different methods used for folate concentration
determination (UVabsorbance and fluorescence) and possibly, while less
likely, by the difference in binding of GNMT with 5-CH
3
-THF-mono and
pentaglutamates.
In the crystal structures of the GNMTfolate complex two folate mole-
cules in the tetrameric structure of GNMT were identified as shown in
Fig. 11.4.
An extraordinary feature of the folate-binding sites in the GNMT struc-
ture is that all four subunits participate in interaction with each folate
molecule. As it is shown in Fig. 11.4, one of folates interacts with 5 N-
terminal residues of subunits Aand Cand with fragments of 206217 residues
fromsubunits Band D. Another folate molecule interacts with 5 N-terminal
residues of subunits B and D and residues 206217 from subunits A and C.
The crystal structure of that complex provides the clue to on how folate
inhibits GNMT activity. From GNMT inhibition studies of folate (Yeo
et al., 1999), it is known that folate inhibition is noncompetitive regarding
both substrates, that is, folate does not compete for binding sites for any of
the substrates. The crystal structure of GNMTfolate complex confirmed
that earlier finding.
In the mechanism of GNMT enzymatic reaction proposed by Takata
et al. (2003), the first step is the binding of AdoMet. In the apoprotein,
however, the entrances to active centers on each subunit are closed by
U-loops formed by N-terminal residues of the adjacent subunits. Moreover,
the 5 N-terminal residues of all subunits interact as is shown in Fig. 11.4.
When AdoMet enters the active center, it displaces the N-terminal U-loops
and disrupts N-terminal residues interactions. The crystal structure of
GNMT with AdoMet showed that in the GNMTAdoMet complex,
N-terminal fragments of all subunits no longer interact and are exposed
GNMT-methyltetrahydrofolate 339
into solvent (Takata et al., 2003). Thus, binding AdoMet requires disruption
of the interaction of all N-terminal fragments of GNMT subunits.
That mechanism implies that any modification of the interaction of
N-terminal fragments will result in activation/inhibition of the enzymatic
reaction. Bound folate molecules strongly interact with the first 7 N-terminal
residues of each subunit. As a result, their flexibility significantly decreased
and much higher concentration of AdoMet is needed to enter the active
centers, that is, the enzymatic reaction is inhibited by folate.
VI. Conclusions
Cell and organism homeostasis requires strict regulation of all bio-
chemical processes in response to changes of environment. GNMT is one of
the most important parts of that regulation machinery. Therefore, exact
Figure 11.4 Crystal structure of rat glycine-N-methyltransferase (GNMT)5-methyl-
tetrahydrofolate (5-CH
3
-THF) complex. The individual subunits are in black and in
light gray and marked as A, B, C, and D with interacting N-terminal as N
A
, N
B
, N
C
,
and N
D
, respectively. Two 5-CH
3
-THF monoglutamate molecules are shown in
sphere mode and marked by arrows. Figure was drawn by using atom coordinates
taken from Protein Data Bank accession code 2IDK.
340 Zigmund Luka
knowledge of how that part is built and how it operates is needed for a
correct explanation of any abnormality in the cell and correct prediction of
how that affects the cells and organism.
The data discussed in this chapter show that knowledge of the role of
GNMT as methylating enzyme and folate-binding protein allows us to
predict the major effect of inactivation of that enzyme: increase in the
level of methionine and AdoMet, folate binding and physiological conse-
quences for liver. Our knowledge of GNMT function, however, is still not
sufficient to answer other questions. We still do not know why GNMT is
abundant in pancreas, kidney, and prostate but is not expressed in other
tissues. Is that related to secretion by pancreas and prostate? If so, what is the
mechanism of GNMT participation in that process? Another question is
how different forms of folate are distributed among all folate-binding
protein in the cell and how it is related to any abnormality, including
diseases? This is also a practical problem since folate derivatives are widely
used in medicine. Another unanswered question is, what is the primary
reason for the GNMT gene being completely repressed in tumor cells. Is it
just a trivial consequence of a shift of the total pattern of gene expression or
is repression of GNMT, a part of a tumor growth triggering mechanism?
The hope is that recent development in GNMT studies will bring the
answers to most of these questions.
ACKNOWLEDGMENTS
This work was supported by Grant DK15289 from the National Institutes of Health. The
author thanks Prof. Conrad Wagner for his critical reading of the manuscript.
REFERENCES
Aida, K., Tawata, M., Negishi, M., and Onaya, T. (1997). Mouse glycine N-methyltransferase
is sexually dimorphic and regulated by growth hormone. Horm. Metab. Res. 29, 646649.
Augoustides-Savvopoulou, P., Luka, Z., Karyda, S., Stabler, S. P., Allen, R. H.,
Patsiaoura, K., Wagner, C., and Mudd, S. H. (2003). Glycine N methyltransferase
deficiency, a new patient with a novel mutation. J. Inherit. Metab. Dis. 26, 745759.
Appling, D. R. (1991). Compartmentation of folate-mediated one-carbon metabolism in
eukaryotes. FASEB J. 5, 26452651.
Balaghi, M., Horne, D. W., and Wagner, C. (1993). Hepatic one-carbon metabolism in
early folate deficiency in rats. Biochem. J. 291, 145149.
Blumenstein, J., and Williams, G. R. (1960). The enzymatic N-methylation of glycine.
Biochem. Biophys. Res. Comm. 3, 259263.
Brown-Borg, H. M., Rakoczy, S. G., and Uthus, E. O. (2005). Growth hormone alters
methionine and glutathione metabolism in Ames dwarf mice. Mech. Ageing Dev. 126,
389398.
GNMT-methyltetrahydrofolate 341
Chen, Y. M., Shiu, J. Y., Tzeng, S. J., Shih, L. S., Chen, Y. J., Lui, W. Y., and Chen, P. H.
(1998). Characterization of glycine-N-methyltransferase-gene expression in human
hepatocellular carcinoma. Int. J. Cancer 75, 787793.
Chen, Y. M., Chen, L. Y., Wong, F. H., Lee, C. M., Chang, T. J., and Yang-Feng, T. L.
(2000). Genomic structure, expression, and chromosomal localization of the human
glycine N-methyltransferase gene. Genomics 66, 4347.
Chen, Z., Karaplis, A. C., Ackerman, S. L., Pogribny, I. P., Melnyk, S., Lussier-Cacan, S.,
Chen, M. F., Pai, A., John, S. W., Smith, R. S., Bottiglieri, T., Bagley, P., et al. (2001).
Mice deficient in methylenetetrahydrofolate reductase exhibit hyperhomocysteinemia
and decreased methylation capacity, with neuropathology and aortic lipid deposition.
Hum. Mol. Genet. 10, 433443.
Clarke, S., and Banfild, K. (2001). S-adenosylmethionine dependent methyltransferase.
In Homocysteine in Health and Disease (R. Carmel and D. W. Jacobsen, eds.),
pp. 6378. Cambridge University Press, Cambridge.
Clifford, A. J., Heid, M. K., Muller, H. G., and Bills, N. D. (1990). Tissue distribution and
prediction of total body folate of rats. J. Nutr. 120, 16331639.
Cook, R. J. (2001). Folate metabolism. In Homocysteine in Health and Disease (R. Carmel
and D. W. Jacobsen, eds.), pp. 113134. Cambridge University Press, Cambridge.
Cook, R. J., and Wagner, C. (1981). Measurement of a folate binding protein from rat liver
cytosol by radioimmunoassay. Arch. Biochem. Biophys. 208, 358364.
Cook, R. J., and Wagner, C. (1984). Glycine N-methyltransferase is a folate binding protein
of rat liver cytosol. Proc. Natl. Acad. Sci. USA 84, 36313634.
Foo, S. K., McSloy, R. M., Rousseau, C., and Shane, B. (1982). Folate derivatives in human
cells: Studies on normal and 5,10-methylenetetrahydrofolate reductase-deficient fibroblasts.
J. Nutr. 112, 16001608.
Fu, Z., Hu, Y., Konishi, K., Takata, Y., Ogawa, H., Gomi, T., Fujioka, M., and
Takusagawa, F. (1996). Crystal structure of glycine N-methyltransferase from rat liver.
Biochemistry 35, 1198511993.
Fujioka, M., Takata, Y., Konishi, K., and Ogawa, H. (1987). Function and reactivity of
sulfhydryl groups of rat liver glycine methyltransferase. Biochemistry 26, 696702.
Halsted, C. H. (1979). The intestinal absorption of folates. Am. J. Clin. Nutr. 32, 846855.
Heady, J. E., and Kerr, S. J. (1973). Purification and characterization of glycine
N-methyltransferase. J. Biol. Chem. 248, 6972.
Heady, J. E., and Kerr, S. J. (1975). Alteration of glycine N-methyltransferase activity in
fetal, adult, and tumor tissues. Cancer Res. 35, 640643.
Henderson, G. B. (1990). Folate-binding proteins. Annu. Rev. Nutr. 10, 319335.
Herbert, V., and Zalusky, R. (1962). Interrelations of vitamin B12 and folic acid metabolism:
Folic acid clearance studies. J. Clin. Invest. 41, 12631276.
Huang, Y., Komoto, J., Konishi, K., Takata, Y., Ogawa, H., Gomi, T., Fujioka, M., and
Takusagawa, F. (2000). Mechanisms for auto-inhibition and forced product release in
glycine N-methyltransferase, crystal structures of wild-type, mutant R175K and
S-adenosylhomocysteine-bound R175K enzymes. J. Mol. Biol. 298, 149162.
Jacobs, R. L., Stead, L. M., Brosnan, M. E., and Brosnan, J. T. (2001). Hyperglucagonemia
in rats results in decreased plasma homocysteine and increased flux through the transsul-
furation pathway in liver. J. Biol. Chem. 276, 4374043747.
Kim, D. W., Huang, T., Schirch, D., and Schirch, V. (1996). Properties of tetrahydropter-
oylpentaglutamate bound to 10-formyltetrahydrofolate dehydrogenase. Biochemistry 35,
1577215783.
Kloor, D., Karnahl, K., and Kompf, J. (2004). Characterization of glycine
N-methyltransferase from rabbit liver. Biochem. Cell Biol. 82, 369374.
Konishi, K., and Fujioka, M. (1987). Chemical modification of a functional arginine residue
of rat liver glycine methyltransferase. Biochemistry 26, 84968502.
342 Zigmund Luka
Konishi, K., and Fujioka, M. (1988). Rat liver glycine methyltransferase. Cooperative
binding of S-adenosylmethionine and loss of cooperativity by removal of a short
NH
2
-terminal segment. J. Biol. Chem. 263, 1338113385.
Liang, C. R., Leow, C. K., Neo, J. C., Tan, G. S., Lo, S. L., Lim, J. W., Seow, T. K.,
Lai, P. B., and Chung, M. C. (2005). Proteome analysis of human hepatocellular carcinoma
tissues by two-dimensional difference gel electrophoresis and mass spectrometry. Proteomics
5, 22582271.
Liu, H. H., Chen, K. H., Shih, Y. P., Lui, W. Y., Wong, F. H., and Chen, Y. M. (2003).
Characterization of reduced expression of glycine N-methyltransferase in cancerous hepatic
tissues using two newly developed monoclonal antibodies. J. Biomed. Sci. 10, 8797.
Luka, Z., and Wagner, C. (2003a). Expression and purification of glycine
N-methyltransferases in Escherichia coli. Protein Expr. Purif. 28, 280286.
Luka, Z., and Wagner, C. (2003b). Human glycine N-methyltransferase is unfolded by urea
through a compact monomer state. Arch. Biochem. Biophys. 420, 153160.
Luka, Z., and Wagner, C. (2003c). Effect of naturally occurring mutations in human glycine
N-methyltransferase on activity and conformation. Biochem. Biophys. Res. Commun. 312,
10671072.
Luka, Z., Cerone, R., Phillips, J. A., 3rd., Mudd, H. S., and Wagner, C. (2002). Mutations
in human glycine N-methyltransferase give insights into its role in methionine metabo-
lism. Hum. Genet. 110, 6874.
Luka, Z., Ham, A.-J., Norris, J. L., Yeo, E.-J., Yermalitsky, V., Glenn, B., Caprioli, R. M.,
Liebler, D. C., and Wagner, C. (2006a). Identification of phosphorylation sites in glycine
N-methyltransferase from rat liver. Protein Sci. 15, 785789.
Luka, Z., Capdevila, A., Mato, J. M., and Wagner, C. (2006b). A glycine
N-methyltransferase knockout mouse model for humans with deficiency of this enzyme.
Transgenic Res. 15, 393397.
Luka, Z., Pakhomova, S., Loukachevitch, L. V., Egli, M., Newcomer, M. E., and
Wagner, C. (2007). 5-methyltetrahydrofolate is bound in intersubunit areas of rat liver
folate-binding protein glycine N-methyltransferase. J. Biol. Chem. 282, 40694075.
Matherly, L. H., and Goldman, D. (2003). Membrane transport of folates. In Vitamines and
Hormones (G. Litwack, ed.), Vol. 66, pp. 405456. Academic Press, California.
Mays, L. L., Borek, E., and Finch, C. E. (1973). Glycine N-methyltransferase is a regulatory
enzyme which increases in ageing animals. Nature 243, 411413.
McKillop, D. J., McNulty, H., Scott, J. M., McPartlin, J. M., Strain, J. J., Bradbury, I.,
Girvan, J., Hoey, L., McCreedy, R., Alexander, J., Patterson, B. K., Hannon-
Fletcher, M., et al. (2006). The rate of intestinal absorption of natural food folates is not
related to the extent of folate conjugation. Am. J. Clin. Nutr. 84, 167173.
McMullen, M. H., Rowling, M. J., Ozias, M. K., and Schalinske, K. L. (2002). Activation
and induction of glycine N-methyltransferase by retinoids are tissue- and gender-specific.
Arch. Biochem. Biophys. 401, 7380.
Min, H., Shane, B., and Stokstad, E. L. (1988). Identification of 10-formyltetrahydrofolate
dehydrogenase-hydrolase as a major folate binding protein in liver cytosol. Biochim.
Biophys. Acta 967, 348353.
Moller, M. T., Samari, H. R., Fengsrud, M., Stromhaug, P. E., oStvold, A. C., and
Seglen, P. O. (2003). Okadaic acid-induced, naringin-sensitive phosphorylation of
glycine N-methyltransferase in isolated rat hepatocytes. Biochem. J. 373, 505513.
Mudd, S. H., Cerone, R., Schiaffino, M. C., Fantasia, A. R., Minniti, G., Caruso, U.,
Lorini, R., Watkins, D., Matiaszuk, N., Rosenblatt, D. S., Schwahn, B., Rozen, R., et al.
(2001). Glycine N-methyltransferase deficiency, a novel inborn error causing persistent
isolated hypermethioninaemia. J. Inherit. Metab. Dis. 24, 448464.
Nieman, K. M., Hartz, C. S., Szegedi, S. S., Garrow, T. A., Sparks, J. D., and
Schalinske, K. L. (2006). Folate status modulates the induction of hepatic glycine
GNMT-methyltetrahydrofolate 343
N-methyltransferase and homocysteine metabolism in diabetic rats. Am. J. Physiol.
Endocrinol. Metab. 291, 12351242.
Ogawa, H., and Fujioka, M. (1982a). Purification and properties of glycine
N-methyltransferase from rat liver. J. Biol. Chem. 257, 34473452.
Ogawa, H., and Fujioka, M. (1982b). Induction of rat liver glycine methyltransferase by high
methionine diet. Biochem. Biophys. Res. Commun. 108, 227232.
Ogawa, H., Gomi, T., Horii, T., Ogawa, H., and Fujioka, M. (1984). Molecular cloning of
cDNA for rat glycine methyltransferase. Biochem. Biophys. Res. Commun. 124, 4450.
Ogawa, H., Konishi, K., Takata, Y., Nakashima, H., and Fujioka, M. (1987). Rat glycine
methyltransferase. Complete amino acid sequence deduced from a cDNA clone and
characterization of the genomic DNA. Eur. J. Biochem. 168, 141151.
Ogawa, H., Gomi, T., and Fujioka, M. (1993). Mammalian glycine N-methyltransferases.
Comparative kinetic and structural properties of the enzymes from human, rat, rabbit and
pig livers. Comp. Biochem. Physiol. B. 106, 601611.
Ogawa, H., Gomi, T., Takata, Y., Date, T., and Fujioka, M. (1997). Recombinant expression
of rat glycine N-methyltransferase and evidence for contribution of N-terminal acetylation
to co-operative binding of S-adenosylmethionine. Biochem. J. 327, 407412.
Pacha, I. (2000). Development if intestinal transport function in mammals. Physiol. Rev. 80,
16331667.
Pakhomova, S., Luka, Z., Grohmann, S., Wagner, C., and Newcomer, M. E. (2004).
Glycine N-methyltransferases, a comparison of the crystal structures and kinetic proper-
ties of recombinant human, mouse and rat enzymes. Proteins 57, 331337.
Pattanayek, R., Newcomer, M. E., and Wagner, C. (1998). Crystal structure of apo-glycine
N-methyltransferase (GNMT). Protein Sci. 7, 13261331.
Paukert, J. L., Straus, L. D., and Rabinowitz, J. C. (1976). Formyl-methyl-methylenete-
trahydrofolate synthetase-(combined). An ovine protein with multiple catalytic activities.
J. Biol. Chem. 251, 51045111.
Rowling, M. J., and Schalinske, K. L. (2003). Retinoic acid and glucocorticoid treatment
induce hepatic glycine N-methyltransferase and lower plasma homocysteine concentra-
tions in rats and rat hepatoma cells. J. Nutr. 133, 33923398.
Rowling, M. J., McMullen, M. H., Chipman, D. C., and Schalinske, K. L. (2002). Hepatic
glycine N-methyltransferase is up-regulated by excess dietary methionine in rats. J. Nutr.
132, 25452550.
Said, H. M. (2004). Recent advances in carrier-mediated intestinal absorption of water-
soluble vitamins. Ann. Rev. Physiol. 66, 419446.
Said, H. M., and Mohammed, Z. M. (2006). Intestinal absorption of water-soluble vitamins:
An apdate. Curr. Opin. Gastroenterol. 22, 140146.
Schirch, L. V., and Gross, T. (1968). Serine Transhydroxymethylase. Identification as the
threonine and allothreonine aldolases. J. Biol. Chem. 243, 56515655.
Seyoum, E., and Selhub, J. (1998). Properties of food folates determined by stability and
susceptibility to intestinal pteroylpolyglutamate hydrolase action. J. Nutr. 128, 19561960.
Shane, B. (1995). Folate Chemistry and metabolism. In Folate in Health and Disease
(L. B. Bayley, ed.), pp. 122. Marcel Dekker, Inc., New York.
Shane, B., and Stokstad, E. L. (1985). Vitamin B12-folate interrelationships. Annu. Rev.
Nutr. 5, 115141.
Shin, Y. S., Williams, M. A., and Stokstad, E. L. (1972). Identification of folic acid
compounds in rat liver. Biochem. Biophys. Res. Commun. 47, 3543.
Sirotnak, F. M., and Tolner, B. (1999). Carrier-mediated membrane transport of folates in
mammalian cells. Annu. Rev. Nutr. 19, 92122.
Soriano, A., Castillo, R., Christov, C., Andres, J., Moliner, V., and Tunon, I. (2006).
Catalysis in glycine N-methyltransferase: Testing the electrostatic stabilization and
compression hypothesis. Biochemistry 45, 1491714925.
344 Zigmund Luka
Suh, J. R., Herbig, A. K., and Stover, P. J. (2001). New perspectives on folate catabolism.
Annu. Rev. Nutr. 21, 255282.
Suzuki, N., and Wagner, C. (1980). Purification and characterization of a folate binding
protein from rat liver cytosol. Arch. Biochem. Biophys. 199, 236248.
Swanson, D. A., Liu, M. L., Baker, P. J., Garrett, L., Stitzel, M., Wu, J., Harris, M.,
Banerjee, R., Shane, B., and Brody, L. C. (2001). Targeted disruption of the methionine
synthase gene in mice. Mol. Cell. Biol. 21, 10581065.
Takata, Y., Huang, Y., Komoto, J., Yamada, T., Konishi, K., Ogawa, H., Gomi, T.,
Fujioka, M., and Takusagawa, F. (2003). Catalytic mechanism of glycine
N-methyltransferase. Biochemistry 42, 83948402.
Tseng, T. L., Shih, Y. P., Huang, Y. C., Wang, C. K., Chen, P. H., Chang, J. G.,
Yeh, K. T., Chen, Y. M., and Buetow, K. H. (2003). Genotypic and phenotypic
characterization of a putative tumor susceptibility gene, GNMT, in liver cancer. Cancer
Res. 63, 647654.
Uthus, E. O., and Brown-Borg, H. M. (2003). Altered methionine metabolism in long
living Ames dwarf mice. Exp. Gerontol. 38, 491498.
Velichkova, P., and Himo, F. (2005). Methyl transfer in glycine N-methyltransferase.
A theoretical study. J. Phys. Chem. B. 109, 82168219.
Villanueva, J. A., and Halsted, C. H. (2004). Hepatic transmethylation reactions in micropigs
with alcoholic liver disease. Hepatology 39, 13031310.
Waditee, R., Tanaka, Y., Aoki, K., Hibino, T., Jikuya, H., Takano, J., Takabe, T., and
Takabe, T. (2003). Isolation and functional characterization of N-methyltransferases
that catalyze betaine synthesis from glycine in a halotolerant photosynthetic organism
Aphanothece halophytica. J. Biol. Chem. 278, 49324942.
Wagner, C. (1982). Cellular folate bindng proteins; function and significance. Ann. Rev.
Nutr. 2, 229248.
Wagner, C. (1995). Biochemical role of folate in cellular metabolism. In Folate in Health
and Disease (L. B. Bayley, ed.), pp. 2342. Marcel Dekker Inc., New York.
Wagner, C., Briggs, W. T., and Cook, R. J. (1985). Inhibition of glycine
N-methyltransferase activity by folate derivatives, implications for regulation of methyl
group metabolism. Biochem. Biophys. Res. Commun. 127, 746752.
Wagner, C., Decha-Umphai, W., and Corbin, J. (1989). Phosphorylation modulates the
activity of glycine N-methyltransferase, a folate binding protein. In vitro phosphorylation
is inhibited by the natural folate ligand. J. Biol. Chem. 264, 96389642.
Williams, K. T., and Schalinske, K. L. (2007). New insights into the regulation of methyl
group and homocysteine metabolism. J. Nutr. 137, 311314.
Wittwer, A. J., and Wagner, C. (1980). Identification of folate binding protein of mitochon-
dria as dimethylglycine dehydrogenase. Proc. Natl. Acad. Sci. USA 77, 44844488.
Xu, G. P., and Snoswell, A. M. (1985). Disturbance of methyl group metabolism in alloxan-
diabetic sheep. Biochem. Int. 10, 897905.
Yeo, E. J., and Wagner, C. (1992). Purification and properties of pancreatic glycine
N-methyltransferase. J. Biol. Chem. 267, 2466924674.
Yeo, E. J., and Wagner, C. (1994). Tissue distribution of glycine N-methyltransferase,
a major folate-binding protein of liver. Proc. Natl. Acad. Sci. USA 91, 210214.
Yeo, E. J., Briggs, W. T., and Wagner, C. (1999). Inhibition of glycine N-methyltransferase
by 5-methyltetrahydrofolate pentaglutamate. J. Biol. Chem. 274, 3755937564.
Zamierowski, M. M., and Wagner, C. (1977). Identification of folate binding proteins in rat
liver. J. Biol. Chem. 252, 933938.
GNMT-methyltetrahydrofolate 345
C H A P T E R T W E L V E
Mechanism-Based Inhibitors of
Folylpoly-g-Glutamate Synthetase and
g-Glutamyl Hydrolase: Control of
Folylpoly-g-Glutamate Homeostasis
as a Drug Target
James K. Coward* and John J. McGuire
Contents
I. Introduction 348
II. Folylpoly-g-Glutamate Homeostasis as a Drug Target 350
A. FPGS and GH: Biochemistry 350
B. FPGS and GH: Structure 352
III. Design of Fluoroglutamate-Containing Folates and Antifolates as
FPGS or GH Alternate Substrates and/or Inhibitors 354
A. Replacement of hydrogen by fluorine: Electronics
versus sterics 354
B. Predicted effect of fluorine substitution in free amino acids and
in fluoroglutamate-containing oligo-g-glutamates 356
IV. Synthesis of Fluorine-Containing Folates and Antifolates from the
Corresponding Fluoroglutamates and Related Fluoroamino Acids 356
A. 4-Fluoroglutamic acid 356
B. 4,4-Difluoroglutamic acid and derivatives 357
C. 3,3-Difluoroglutamic acid 358
D. Folates and antifolates derived from 4-FGlu 358
E. Folates and antifolates derived from 4,4-F
2
Glu and 4,4-F
2
Orn 359
F. Folates and antifolates derived from 3,3-F
2
Glu 360
G. Fluoroglutamate derivatives: Summary and conclusions 361
V. Design of Phosphorus-Containing Pseudopeptides as
FPGS Inhibitors 361
A. Introduction 361
Vitamins and Hormones, Volume 79
#
2008 Elsevier Inc.
ISSN 0083-6729, DOI: 10.1016/S0083-6729(08)00412-3 All rights reserved.
* Departments of Medicinal Chemistry and Chemistry, University of Michigan, 3813 Chemistry, 930 N.
University, Ann Arbor, Michigan 48109
{
Grace Cancer Drug Center, Roswell Park Cancer Institute, Buffalo, New York 14263
347
B. Synthesis of phosphorus-containing pseudopeptide inhibitors
of FPGS 362
C. Biochemical properties of phosphorus-containing
pseudopeptide analogues of g-glutamyl conjugates of folate
and antifolates 364
D. Phosphorus-containing pseudopeptides: Summary
and conclusions 365
VI. Design of Epoxide-Containing Peptidomimetics as GH Inhibitors 366
A. Introduction 366
B. Synthesis of epoxide-containing peptidomimetics 366
VII. Conclusions 366
Acknowledgment 367
References 367
Abstract
Intracellular folate pools consist primarily of g-glutamyl isopeptide conjugates of
reduced forms of the vitamin, folic acid. Biosynthesis of these oligomeric isopep-
tides is catalyzed by the enzyme folylpoly-g-glutamate synthetase (FPGS).
The highly anionic character of the oligomers renders them unable to cross
the cell membrane and, therefore, these forms of reduced folates (and certain
antifolates) accumulate in cells to high concentration. g-Glutamyl hydrolase
(GH) catalyzes the hydrolysis of the oligo-g-glutamates derivatives to monoglu-
tamyl forms, which are substrates for the reduced folate carrier and able to
exit the cell. This chapter describes research, primarily from our laboratories,
on the design, synthesis, and biochemical evaluation of several novel
analogues of glutamic acid, g-glutamyl peptides, and derivatives of folic
acid as well as of antifolate drugs. These include a series of fluoroglutamic
acids, fluoroglutamate-containing isopeptides, phosphorus-containing pseudo-
peptides, and epoxide-containing peptidomimetics. The fluoroglutamic acids and
fluoroglutamate-containing folates and antifolates exhibit position-dependent
effects on the reactions catalyzed by FPGS and GH, thus providing insight into
the catalytic mechanism and control of these enzymes. The phosphinic acid-
containing pseudopeptides are the most potent inhibitors of FPGS identified to
date, and were designed based on mechanistic enzymology data from our
laboratories and others, prior to the publication of any structural information
about the targeted enzymes. 2008 Elsevier Inc.
I. Introduction
Naturally occurring folate conjugates are composed of a series of
oligo-g-glutamate isopeptides attached to pteroic acid via an amide bond
(Fig. 12.1). Although the structures of these compounds have been known
for many years (Stokstad and Koch, 1967), details on their biosynthesis,
348 James K. Coward and John J. McGuire
degradation, and utilization have been elucidated only more recently (Galivan
et al., 2000; McGuire and Coward, 1984; Schirch and Strong, 1989; Shane,
1989). In this chapter, we describe research, primarily from our own labora-
tories, that has sought to exploit our understanding of the reactions catalyzed
by folylpoly-g-glutamate synthetase (FPGS, EC 6.3.2.17) and g-glutamyl
hydrolase (GH, EC 3.4.19.9). These enzymes act in concert to maintain
levels of intracellular folylpoly-g-glutamates sufficient for cell growth, as well
as adequate levels of antifolate chemotherapeutic agents to inhibit the growth
of cells targeted in the treatment of cancer and other pathological conditions
such as rheumatoid arthritis. Early work demonstrated an absolute require-
ment for folylpoly-g-glutamate synthesis for cell viability in Chinese hamster
ovary cells. A mutant cell line, AUXB1, that is auxotrophic for glycine,
adenosine, and thymidine (McBurney and Whitmore, 1974), or glycine
and methionine (Taylor and Hanna, 1977), contains primarily low levels of
monoglutamate derivatives of reduced folates, and it was concluded that the
defect in this cell line as well as a temperature-sensitive variant, tsAUXB1, is
the enzyme responsible for linking glutamate residues onto intracellular folate
derivatives. Taylor and Hanna (1977) subsequently identified FPGS as the
single enzyme responsible for the multiple auxotrophy. Two consequences
HN
O
O
O
O
CO
2
CO
2
CO
2
CO
2
CO
2
CO
2
CO
2
CO
2
H
2
N
H
2
N
H
2
N
N
H
N
H
N
H
N
H
N
H
N
H
N
H
H
N
N
H
N
H
N
H
N
H
H
N
H
N
HN
HN
O
O
O
O
g GH FPGS
n
Figure 12.1 Effects of FPGS-catalyzed glutamate ligation and GH-catalyzed isopeptide
hydrolysis on folate efflux and uptake in intact cells.
FPGS and GH Inhibitors 349
of folylpolyglutamate formation were clear from the early work on AUXB1
mutants, namely the importance of polyglutamates in one-carbon biochem-
istry (Schirch and Strong, 1989) and in folate retention (Matherly and
Goldman, 2003) (Fig. 12.1). Associated with the impact of polyglutamyla-
tion, resistance to antifolates in cancer chemotherapy has been attributed to
decreased expression of the FPGS gene that results in decreased formation of
poly-g-glutamyl derivatives of antifolate drugs and more rapid efflux from
the cell (Pizzorno et al., 1988). With some drugs, for example, Tomudex,
decreased polyglutamylation also decreases drug target inhibitory potency
(Clarke et al., 2000). Natural variations in the gene coding for GH arising
from single nucleotide polymorphisms (Chave et al., 2003) lead to increased
expression of GH, which could lead to resistance to antifolate drugs such as
methotrexate (MTX) (Schneider and Ryan, 2006). Thus, it is likely that
development of small molecule inhibitors of FPGS (McGuire and Coward,
2003; McGuire et al., 1996b) and/or GH (Galivan et al., 2000) could have
clinical application (Longo et al., 1997; Rots et al., 1999)
II. Folylpoly-g-Glutamate Homeostasis as a
Drug Target
A. FPGS and GH: Biochemistry
The reactions catalyzedbyFPGSandGHare showninSchemes 12.1and12.2.
The carbon atom involved either in peptide bond formation (FPGS) or in
hydrolysis (GH) is converted from a carbonyl group (sp
2
) to an unstable
tetrahedral intermediate (sp
3
), which collapses to expel either an inorganic
phosphate (FPGS) or a glutamate leaving group (GH). In the case of FPGS,
activation of the g-carboxyl group involves the ATP-dependent formation
of a g-glutamyl phosphate (Banerjee et al., 1988). In the case of GH, an
active site cysteine (Cys110 in the human enzyme) has been shown to be
essential for enzyme activity (Chave et al., 1999, 2000) and is presumed to
be involved in formation of an acyl enzyme intermediate (Alexander et al.,
2008). Given that the folate conjugates under discussion in this chapter
contain oligomers of glutamic acid, FPGS can be considered a polymerase
while GH can be compared with other enzymes that hydrolyze polymeric
biomolecules, that is, nucleases (nucleic acids) or glycohydrolases (oligosac-
charides). As such, it is important to determine if each reaction proceeds via
either a processive or a distributive mechanism (McClure and Chow, 1980).
Analysis of FPGS-catalyzed ligation of glutamic acid to either (6RS) 5,6,7,
8-tetrahydrofolate (H
4
PteGlu, rat FPGS) (McGuire et al., 1980; Taylor
and Hanna, 1977) or (6R) 5,10-dideazatetrahydrofolate (DDAH
4
PteGlu,
human FPGS) (Tomsho et al., 2005) indicated that product chain length is
inversely dependent on substrate concentration. As suggested by these
350 James K. Coward and John J. McGuire
data, a processive mechanism of multiple glutamate ligations has been
demonstrated at low concentration of DDAH
4
PteGlu (Tomsho et al.,
2008). Similarly, kinetic analysis of GH-catalyzed hydrolysis of a derivative
of Glu-g-Glu revealed that formation of the acyl enzyme intermediate is not
rate-limiting, thus suggesting that the product release may be the rate-
limiting step, a conclusion that is consistent with GH-catalyzing processive
hydrolysis (Alexander et al., 2008). As will be discussed below, these findings
place constraints on the design of selective inhibitors for these enzymes.
C
O
N
H
H
2
N
CO
2
-
CO
2
-
-
O
=
O
3
P
CO
2
- C
O
N
H
CO
2
-
CO
2
-
CO
2
-
ATP
ADP
C
O
N
H
P
CO
2
-
CO
2
-
-
O
=
O
3
P
CO
2
-
FPGS
(X = CH
2
)
Mechanism-Based
FPGS inhibitors
Het
Het
O
N
H
CO
2
-
CO
2
-
ATP
ADP
FPGS
O
N
H
C
O
O
PO
3
=
X=NH, O, CH
2
H
2
N
CO
2
-
CO
2
-
Z
Z
Het
C
O
N
H
CO
2
-
H
N CO
2
-
O CO
2
-
Z
Donor substrate
Acceptor substrate
g-Glutamyl isopeptide product of FPGS-catalyzed ligation
involving fluoroglutamate-containing acceptor or donor
substrates (Y=H or F, Z=H or F).
Postulated phosphorylated phosphinate-containing
pseudopeptide resulting from FPGS-catalyzed, ATP-
dependent reaction
P
O
X
Y
Y
Y
Y
Folate-based
P
i
Het =
N
N
HN
N
O
H
2
N
N
H
N
N
N
N H
2
N
NH
2
N
Me
N
H
HN
N
O
H
2
N
C
H
2
C
C
O
+
Methotrexate-based Lometrexol-based
O
O
CO
2
-
Scheme 12.1 Mechanism of FPGS-catalyzed glutamate ligation, effects of fluoroglu-
tamate-containing acceptor or donor substrates, and inhibition by phosphorous-
containing pseudopeptide analogues.
FPGS and GH Inhibitors 351
B. FPGS and GH: Structure
Crystal structures of both enzymes have been reported in the literature.
FPGS from Lactobacillus casei has been crystallized as a binary complex
(2.4 A
H
N CO
2
O
CO
2
Z
GH-Cys
110
-SH
C
O
N
H
CO
2
H
N CO
2
CO
2
S O
Cys
110
GH
H
2
N
CO
2
CO
2
Z
C
O
N
H
CO
2
S
Cys
110
O
Z
Y
Y
Y
C
O
N
H
CO
2
O
Y
GH
GH-Cys
110
-SH
H
2
O
CO
2
CO
2
O
CO
2
GH-Cys
110
-SH
CO
2
CO
2
CO
2
S
Cys
110
GH
OH
Substituted o- or p-aminobenzoyl glutamate product of
GH-catalyzed hydrolysis involving fluoroglutamate-containing
isopeptide substrates(Y=H or F, Z=H or F).
Postulatead GH inactivation by an
epoxide-containing pseudopeptide
Mechanism-Based
GH Inhibitor
RNH
RNH
RNH
RNH
R = H (o- or p-), CH
3
, Het (see scheme 1)
Scheme 12.2 Mechanism of GH-catalyzed hydrolysis of g-glutamyl isopeptides,
effects of fluoroglutamate-containing isopeptide substrates, and postulated inhibition
by epoxide-containing peptidomimetics.
352 James K. Coward and John J. McGuire
structure of the bifunctional protein FolC from Escherichia coli, which
harbors both FPGS and 7,8-dihydrofolate synthetase [DHFS, Eq. (12.1)]
H
2
PteOH Glu ATP ! H
2
PteGlu ADP 12:1
activities, has been reported (Mathieu et al., 2005). Both the N-terminal and
C-terminal domains of FolC are superimposable with those of the L. casei
enzyme. Most interestingly, the structural data indicate that dihydropteroate
can bind to the C-terminal domain (1.9 A
) versus H
(0.790 A
). However, as shown in Table 12.1, the van der Waals radii and
length of the CX bond are both considerably longer when X F versus X
H; in fact, the CF bond length is more similar to that of a COH bond
than it is to a CH bond. In addition, the van der Waals volume of a CF
3
group is ca. 50% larger than that of a CH
3
group and more similar to that of a
CH
3
CH
2
group (Table 12.1). In agreement with the large difference in
electronegativity between H and F (Table 12.1), the effects of fluorine
substitution on pK
a
values of the three acidic groups in glutamic acid,
a-NH
3
, a-CO
2
H, and g-CO
2
H are substantial (Table 12.2). For the
purposes of this review, the effects of interest are those on the g-CO
2
H and
a-NH
3
Sterics
Atomic radius (A
)
a
H 0.790
F 0.570
van der Waals radius (A
)
b
H 1.20
F 1.47
Bond length (A
, R
3
CX)
b
H 1.099
F 1.428
OH 1.440
van der Waals volume (A
3
/substituent)
c
CH
3
21.6
CF
3
39.8
CH
3
CH
2
38.9
(CH
3
)
2
CH 56.2
Electronegativity (w)
b
H 2.20
F 3.98
a
http://environmentalchemistry.com/yogi/periodic/atomicradius.html
b
Timperley and White (2003).
c
Leroux (2004).
Table 12.2 Effect of uorine on pK
a
of glutamic acid
a
H
3
N CO
2
H
CO
2
H
X
X aNH
3
aCO
2
H gCO
2
H
H
b
9.67 2.19 4.25
3-F (8.6) (1.6) (3.8)
3,3-F
2
c
7.45 0.5 3.25
4-F 9.45
d
(2.1) (2.5)
4,4-F
2
(8.5) (1.4) (1.2)
a
Values in parentheses estimated based on known fluorine substituent effects (see
Schlosser, 1998).
b
Jencks and Regenstein (1976).
c
McGuire et al. (1990).
d
Unkeless and Goldman (1971).
FPGS and GH Inhibitors 355
of FPGS and GHhave been published (Coward et al., 1991; Tsukamoto et al.,
1996a) and serve as a foundation for the material reviewed in this chapter.
B. Predicted effect of fluorine substitution in free amino acids
and in fluoroglutamate-containing oligo-g-glutamates
In the FPGS reaction, the ability of the folate/antifolate analogue to form
the g-glutamyl phosphate intermediate (Scheme 12.1) would be reduced by
a decrease in pK
a
of the g-CO
2
H as a result of a decreased nucleophilicity of
the conjugate base. Similarly, a decrease in the pK
a
of the a-NH
3
group
due to fluorine substitution would lead to a decrease in nucleophilicity of
the conjugate base, resulting in attenuated reactivity of the fluoroglutamic
acid as an FPGS acceptor substrate.
2
The magnitude of these electronic
effects is obviously dependent on the position of the fluorine substitution on
the glutamate side chain (Table 12.2). In the case of GH-catalyzed hydro-
lysis of g-glutamyl peptides (Scheme 12.2), the effect of fluorine substitution
is also expected to be position-dependent. Thus, based on the known
enhanced electrophilicity of a,a-difluoro ketones (Gelb et al., 1985) and
g,g-difluoroglutamate esters (Tsukamoto and Coward, 1996), a similar
enhanced electrophilicity would be expected to lead to g-glutamyl peptides
such as (4,4-difluoro)glutamyl-g-glutamate with enhanced activity as a GH
substrate. The synthesis of selectively fluorinated glutamic acids was pursued
in order to test these predictions.
IV. Synthesis of Fluorine-Containing Folates
and Antifolates from the Corresponding
Fluoroglutamates and Related Fluoroamino
Acids
A. 4-Fluoroglutamic acid
This nonnatural amino acid was studied initially as either of two racemic
diastereomers [DL-erythro (2S,4R; 2R,4S) and DL-threo (2S,4S; 2R,4R)],
and the racemic threo diastereomer was found to act as a chain-terminating
inhibitor of FPGS (McGuire and Coward, 1985). Subsequently, synthesis of
all four stereoisomers was reported by Hudlicky (1993), followed by a very
2
A decrease in pK
a
of the a-NH
3
Het-(3-F
2
or 4-F)Glu-g-Glu, may
be limiting. This decrease in conformational flexibility during multiple
glutamate ligations is consistent with the conclusion that FPGS catalysis
appears to proceed via a processive mechanism whereby the newly formed
product containing one additional glutamate does not dissociate from the
enzyme prior to the next glutamate ligation reaction (Tomsho et al., 2008).
Finally, although not discussed in detail in this chapter, extensive cell culture
experiments indicate that the overall effects of fluoroglutamate incorpora-
tion into folates and antifolates result in complex and, as yet, incompletely
understood ramifications on folate-dependent one-carbon biochemistry
and antifolate cytotoxicity as a result of altering polyglutamate biosynthesis
and/or hydrolysis (Galivan et al., 1985b; McGuire et al., 1996a; Tsukamoto
et al., 1996b).
V. Design of Phosphorus-Containing
Pseudopeptides as FPGS Inhibitors
A. Introduction
The reactions catalyzed by FPGS and GH proceed via intermediates that
involve unstable tetrahedral intermediates (Schemes 12.1 and 12.2). Our
design of mechanism-based inhibitors of these enzymes has focused on either
mimicking (FPGS) or intercepting (GH) formation of these intermediates.
Much as the research described above has focused on organofluorine chemis-
try, the research described below involves organophosphorus chemistry.
Initially, our efforts were aimed at synthesizing b-ketophosphonate analogue,
1b (R CO
2
, X = CH
2
, Y = O (acyl phosphate)
b, R=CO
2
or H, X=Y=CH
2
(b-keto phosphonate)
c, R=CO
2
, X=NH, Y=CH
2
(b-amido phosphonate)
1
Het
Structure 1 Structure of g-glutamyl phosphate intermediate, 1a, formed during FPGS-
catalyzed glutamate ligation, and structures of two stable analogues, 1b and 1c.
362 James K. Coward and John J. McGuire
ester containing an appended glutaric acid moiety derived ultimately from
L-glutamic acid. This ester could be coupled to an L-glutamic acid diethyl
ester or 2-hydroxyglutaric acid diethyl ester via the corresponding phos-
phonochloridate or a Mitsunobu coupling, respectively, to afford the phos-
phonamidate and phosphonate analogues of the isopeptide, Glu-g-Glu
(Malachowski and Coward, 1994a,b). Subsequent research indicated that
related phosphonamidates are unstable in aqueous media (Chen et al., 1997).
Therefore, further synthetic efforts on FPGS inhibitors of this class focused
on the phosphonate and phosphinate analogues. A modified synthesis of the
phosphonate analogue of Glu-g-Glu involved the use of the phosphonic
acid monomethyl ester derived ultimately from L-homoserine (Nair et al.,
1995) and (2S) 2-hydroxyglutaric acid dibenzyl ester. Further elaboration
led to the MTX-based phosphonate pseudopeptide and the corresponding
p-ABA-based analogue, with defined stereochemistry at both chiral centers
of this pseudopeptide (2
0
S, 2
00
S), for evaluation as inhibitors of FPGS and
GH, respectively (Tsukamoto et al., 1998).
Synthesis of the phosphinic acid analogues proved to be more difficult,
due primarily to the harsh reaction conditions associated with formation of
the C-P-C array required in this class of compounds. Use of trivalent
phosphorus (P
III
) nucleophiles and electrophilic partners via either conjugate
addition or S
N
2 reactions led first to simplified (Malachowski and Coward,
1994a), then more complex (Valiaeva et al., 2001) phosphinate analogues of
Glu-g-Glu. Elaboration of the more complex analogue led to the desired
phosphinate pseudopeptides, albeit as a mixtures of racemic diastereomers,
for evaluation as inhibitors of FPGS and GH(Valiaeva et al., 2001). Although
successful in providing the desired phosphinate analogue for biochemical
evaluation (vide infra), this synthesis suffered from being stereorandom.
Astereoselective synthesis of these compounds has been developed that allows
for the synthesis of two diastereomeric (2
0
S, 2
00
S and 2
0
S, 2
00
R) phosphinate
pseudopeptide analogues of Glu-g-Glu (Bartley and Coward, 2005). Signifi-
cant improvements in this synthesis included the use of less harsh conditions to
form the N-terminal PC bond (Deprele and Montchamp, 2001), conjugate
addition to 2-methylene glutaric acid esters containing a chiral auxiliary (Liu
et al., 2002), albeit with only modestly stereoselectivity, to form the
C-terminal CP bond, and ultimate separation of the two diastereomeric
pseudopeptides by standard chromatographic methods. Coupling of the
free pseudopeptides to three heterocyclic platforms (Scheme 12.1) employed
substituted benzoyl azides as previously described (Valiaeva et al., 2001).
As is described below, these compounds are extremely potent inhibitors of
isolated human FPGS but are ineffective as inhibitors of polyglutamate
biosynthesis in intact mammalian cells, presumably due to an inability to
cross the cell membrane because of their highly polar structures. In an attempt
FPGS and GH Inhibitors 363
to overcome this significant limitation of this class of compounds, the synthe-
sis of prodrug esters of the phosphinate-containing pseudopeptides was pur-
sued. The pivaloyloxymethyl (POM) ester was chosen because of its
hydrophobic nature, which should lead to improved membrane transport
(Beaumont et al., 2003), as well as its susceptibility to hydrolysis to the parent
drug once inside the cell (Liederer and Borchardt, 2006). A synthesis of a
tris-POM ester prodrug of the MTX-based phosphinate pseudopeptide has
been reported (Feng and Coward, 2006). Unfortunately, facile hydrolysis of a
single POM ester resulted in a disappointing lack of improvement in phar-
macological properties of two sets of POM esters, one being the MTX-based
phosphinate pseudopeptide and a second being the POM ester derivatives of
MTX itself or its g-glutamyl conjugate, in a series of wild-type and MTX
resistant cell lines (Feng and Coward, 2006). Finally, it should be recalled that
FPGS from bacterial sources is known to be a bifunctional protein that
harbors both FPGS and DHFS activities (Mathieu et al., 2005; Salcedo
et al., 2001). A series of aryl phosphinic acids has been designed as potential
inhibitors of the DHFS domain of the bifunctional protein as potential
antibacterial and antimalarial drugs. The first synthesis of aryl phosphinic
acid analogues of folate and several antifolates has been reported (Yang and
Coward, 2007).
C. Biochemical properties of phosphorus-containing
pseudopeptide analogues of g-glutamyl conjugates
of folate and antifolates
As noted above and reported for similar compounds targeted at glutathionyl
spermidine synthetase (Chen et al., 1997), the phosphonamidate derivatives
(Scheme 12.1, X NH) are unstable in aqueous media and were not inves-
tigated further. However, the corresponding phosphonate (Scheme 12.1,
X O) and phosphinate (Scheme 12.1, X CH
2
) pseudopeptides have
been investigated as possible inhibitors of either GH (phosphonate) or FPGS
(phosphonate and phosphinate). Investigation of the phosphonate analogue as a
possible GH inhibitor was predicated on the hypothesis that GH is a Zn
2
-
dependent metallopeptidase (McGuire and Coward, 1984) and, as such, a
potential target for phosphorus-based tetrahedral mimics (Bartlett and
Marlowe, 1983). However, the lack of any inhibitory activity of the MTX-
based phosphonate pseudopeptide, or the analogous p-ABA-based analogue,
argues strongly against this hypothesis (Tsukamoto et al., 1998). In contrast,
these phosphonates are potent (IC
50
15 nM), competitive (K
i
2 nM)
inhibitors of glutamate carboxypeptidase II (GCP II, EC 3.4.17.21),
a zinc-dependent metallopeptidase, also known as N-acetylated a-linked acidic
dipeptidase or prostate-specific membrane antigen (Tsukamoto et al., 2002).
These data together with the independent demonstration that GH is a cysteine
364 James K. Coward and John J. McGuire
peptidase (Chave et al., 1999) indicate that a different approach to the design of
GH-specific inhibitors is required (vide infra).
Examination of the phosphonate- and phosphinate-containing pseudo-
peptides as FPGS inhibitors was more gratifying. The MTX-based phos-
phonate analogue (MTX-phosphonate) is a very potent competitive
inhibitor (K
i
46 nM) of human FPGS and retains the potent inhibitory
activity of MTX against DHFR with a value of IC
50
0.9 nM (Tsukamoto
et al., 1998). The related phosphinate-containing pseudopeptide (MTX-
phosphinate), a mixture of two racemic diastereomers, is an even more
potent competitive inhibitor (K
i
3.1 nM) and also retains inhibitory
activity against DHFR (IC
50
2.1 nM) (McGuire et al., 2003). Evaluation
of the single isomer phosphinates coupled to three heterocyclic platforms
(Scheme 12.1, Het) has demonstrated that one diastereomer, presumably
2
0
S, 2
00
S, exhibits values of IC
50
ca. 10- to 30-fold lower than the other
diastereomer, presumably 2
0
S, 2
00
R. The more potent diastereomers are
competitive inhibitors with values of K
i
15 nM ( J. J. McGuire et al.,
in preparation). The surprisingly small difference in inhibitory activity
between the two diastereomers suggests that the D-configuration at the
C-terminal glutamate is tolerated in these inhibitors, a conclusion consistent
with the conformational flexibility of distal glutamate residues, that is, n > 0
in Het-pAB-Glu-g-[Glu]
n
-g-OH, as discussed above in the context of the
position-dependent effects of 3,3-F
2
Glu.
D. Phosphorus-containing pseudopeptides: Summary
and conclusions
The successful synthesis and biochemical evaluation of a series of phosphorus-
containing pseudopeptides as potential inhibitors of GH and FPGS has
demonstrated that these compounds, while ineffective against isolated GH,
are potent inhibitors of isolated preparations of GCP II and FPGS. Although
not discussed in detail in this chapter, extensive cell culture experiments
indicate that the highly polar phosphorus-containing pseudopeptides do not
inhibit mammalian cell growth (CCRF-CEM) under conditions where MTX
is highly cytotoxic (McGuire and Coward, 2003; Tsukamoto et al., 1998).
As discussed above, the synthesis of prodrug esters is being pursued in an
attempt to overcome the cell membrane permeability of this class of com-
pounds. This approach has been successful with phosphinate-containing mul-
tisubstrate analogues targeting ras protein farnesyltransferase (Manne et al.,
1995; Patel et al., 1995). It is apparent that the phosphinic acid-containing
pseudopeptides described herein are potent inhibitors of FPGS. If improved
transport can be achieved, they have great promise as potential drugs based on
inhibiting the biosynthesis of folate and antifolate polyglutamates.
FPGS and GH Inhibitors 365
VI. Design of Epoxide-Containing
Peptidomimetics as GH Inhibitors
A. Introduction
Incorporation of fluoroglutamic acid into glutamate-containing isopeptides,
either by chemical synthesis or by FPGS-catalyzed ligation, leads to com-
pounds that exhibit a reduced rate of GH-catalyzed hydrolysis (Alexander
et al., 2008; Licato et al., 1990; McGuire et al., 1990). However, the lack of
predictability associated with this approach coupled with the well-knownpoor
bioavailability and metabolic stability of peptide drugs led us to investigate
alternate peptidomimetics as potential GH inhibitors. Our initial approach
involved the use of phosphorus-containing pseudopeptides to target GH, a
putative Zn
2
metallopeptidase but, as noted above, this was unsuccessful and
it is now well established that GH is a cysteine peptidase (Galivan et al., 2000).
More recent synthetic efforts have focused on epoxide-containing peptides
and pseudopeptides.
B. Synthesis of epoxide-containing peptidomimetics
Initially, extension of our research on derivatives of glutamic acid g-semi-
aldehyde as potential GH inhibitors led to a series of C-terminal epoxides.
Although these compounds are GH inhibitors, they act as alternate substrates,
thus complicating the chemical and kinetic analysis of the inhibition process
(Alexander, 2004). More recently, we have developed methods for the
synthesis of pseudopeptides in which a centrally located epoxide moiety has
replaced the scissile peptide bond in the Glu-g-Glu isopeptide (Scheme 12.2).
The synthetic route involves olefination of an N-terminal glutamic acid
g-semialdehyde by a sulfone. The sulfone is appended to a cyclopentene that
acts as a glutaric acid surrogate. The resulting olefin is converted to an
epoxide, which is then elaborated to the desired peptidomimetic
(Alexander, 2004; D. Majumdar et al., in preparation). Biochemical evalua-
tion of these new potential GH inhibitors will be reported in due course.
VII. Conclusions
This chapter summarizes research, primarily from our laboratories,
on the design, synthesis, and biochemical evaluation of several novel ana-
logues of glutamic acid, g-glutamyl peptides, and derivatives of folic acid and
also of antifolate drugs. These include a series of fluoroglutamic acids,
fluoroglutamate-containing isopeptides, phosphorus-containing pseudopep-
tides, and epoxide-containing peptidomimetics. The enzyme targets of these
366 James K. Coward and John J. McGuire
compounds are enzymes that catalyze the biosynthesis (FPGS, Scheme 12.1)
and hydrolysis (GH, Scheme 12.2) of poly-g-glutamate conjugates of folates
and antifolates. The fluoroglutamic acids and fluoroglutamate-containing
folates and antifolates exhibit position-dependent effects on the reactions
catalyzed by FPGS and GH, thus providing insight into the catalytic mecha-
nism and control of these enzymes. The phosphinic acid-containing pseudo-
peptides are the most potent inhibitors of FPGS identified to date, and were
designed based on mechanistic enzymology data from our laboratories and
others, prior to the publication of any structural information about the targeted
enzymes.
ACKNOWLEDGMENT
Research in our laboratories has been supported by grants from the National Cancer Institute
[CA 28097 ( JKC), CA 43500 ( JJM), CA 16056 (RPCI CCSG)]. We thank our students,
postdoctoral associates, and other colleagues whose names are given in the references for
their enthusiastic pursuit of the research described herein.
REFERENCES
Abell, L. M., and Villafranca, J. J. (1991). Investigation of the mechanismof phosphinothricin
inactivation of Escherichia coli glutamine synthetase using rapid quench kinetic techniques.
Biochemistry 30, 61356141.
Ahluwalia, G. S., Grem, J. L., Hao, Z., and Cooney, D. A. (1990). Metabolism and action of
amino acid analog anticancer agents. Pharmacol. Therapeut. 46, 243271.
Alexander, J. P., Ryan, T. J., Ballou, D. P., and Coward, J. K. (2008). g-Glutamyl hydrolase:
Kinetic characterization of isopeptide hydrolysis using fluorogenic substrates. Biochemistry
47, 12281239.
Alexander, M. D. (2004). The synthesis and characterization of g-glutamyl aldehydes and
epoxides designed as inhibitors of the cysteine protease g-glutamyl hydrolase. Ph.D.
Dissertation. University of Michigan, Ann Arbor, MI.
Banerjee, R. V., Shane, B., McGuire, J. J., and Coward, J. K. (1988). Dihydrofolate synthetase
and folylpolyglutamate synthetase: Direct evidence for intervention of acyl phosphate
intermediates. Biochemistry 27, 90629070.
Bartlett, P. A., and Marlowe, C. K. (1983). Phosphonamidates as transition-state analogue
inhibitors of thermolysin. Biochemistry 22, 46184624.
Bartley, D. M., and Coward, J. K. (2005). A stereoselective synthesis of phosphinic acid
phosphapeptides corresponding to glutamyl-g-glutamate and incorporation into potent
inhibitors of folylpoly-g-glutamyl synthetase. J. Org. Chem. 70, 67576774.
Beaumont, K., Webster, R., Gardner, I., and Dack, K. (2003). Design of ester prodrugs to
enhance oral absorption of poorly permeable compounds: Challenges to the discovery
scientist. Curr. Drug Metab. 4, 461485.
Bohm, H. J., Banner, D., Bendels, S., Kansy, M., Kuhn, B., Muller, K., Obst-Sander, U.,
and Stahl, M. (2004). Fluorine in medicinal chemistry. Chem. Biochem. 5, 637643.
Chave, K. J., Galivan, J., and Ryan, T. J. (1999). Site-directed mutagenesis establishes
cysteine-110 as essential for enzyme activity in human g-glutamyl hydrolase. Biochem J.
343, 551555.
FPGS and GH Inhibitors 367
Chave, K. J., Auger, I. E., Galivan, J., and Ryan, T. J. (2000). Molecular modeling and site-
directed mutagenesis define the catalytic motif in human g-glutamyl hydrolase. J. Biol.
Chem. 275, 4036540370.
Chave, K. J., Ryan, T. J., Chmura, S. E., and Galivan, J. (2003). Identification of single
nucleotide polymorphisms in the human g-glutamyl hydrolase gene and characterization
of promoter polymorphisms. Gene 319, 167175.
Chen, S., Lin, C.-H., Kwon, D. S., Walsh, C. T., and Coward, J. K. (1997). Design,
synthesis, and biochemical evaluation of phosphonate and phosphonamidate analogs of
glutathionylspermidine as inhibitors of glutathionylspermidine synthetase/amidase from
Escherichia coli. J. Med. Chem. 40, 38423850.
Clarke, S. J., Beale, P. J., and Rivory, L. P. (2000). Clinical and preclinical pharmacokinetics
of raltitrexed. Clin. Pharmacokin. 39, 429443.
Cook, J. D., Cichowicz, D. J., George, S., Lawler, A., and Shane, B. (1987). Mammalian
folyl-g-glutamate synthetase: In vitro and in vivo metabolism of folates and analogues and
regulation of folate homeostasis. Biochemistry 26, 530539.
Coward, J. K., McGuire, J. J., and Galivan, J. (1991). The influence of fluoro substituents on
the reactivity of carboxylic acids, amides, and peptides in enzyme-catalyzed reactions.
In Selective Fluorination in Organic and Bioorganic Chemistry ( J. T. Welch, ed.),
pp. 196204. American Chemical Society, Washington, DC.
Debart, F., Abes, S., Deglane, G., Moulton, H. M., Clair, P., Gait, M. J., Vasseur, J. J., and
Lebleu, B. (2007). Chemical modifications to improve the cellular uptake of oligonu-
cleotides. Curr. Topics Med. Chem. 7, 727737.
Deprele, S., and Montchamp, J.-L. (2001). Triethylborane-initiated room temperature
radical addition of hypophosphites to olefins: Synthesis of monosubstituted phosphinic
acids and esters. J. Org. Chem. 66, 67456755.
Eisele, L. E., Chave, K. J., Lehning, A. C., and Ryan, T. J. (2006). Characterization of human
g-glutamyl hydrolase in solution demonstrates that the enzyme is a non-dissociating
homodimer. Biochim. Biophys. Acta. 1764, 14791486.
Feng, Y., and Coward, J. K. (2006). Prodrug forms of N-[(4-deoxy-4-amino-10-methyl)
pteroyl]-glutamate-g-[cP(O)(OH)]-glutarate, a potent inhibitor for folylpoly-g-gluta-
mate synthetase: Synthesis and hydrolytic stability. J. Med. Chem. 49, 770788.
Forsch, R., Bader, H., and Rosowsky, A. (1999). Synthesis of L-2-(N-pteroylamino)-3-(N-
phosphonoacetyl)amino-propanoic acid as an analogue of the putative phosphorylated
intermediate in the g-glutamation of folic acid by folylpolyglutamate synthetase. Pteridines
10, 3946.
Galivan, J., Coward, J. K., and McGuire, J. J. (1985a). The effects of D,L-4-fluoroglutamic
acid on the glutamylation of methotrexate by hepatic cells in vitro. Biochem. Pharmacol. 34,
29952997.
Galivan, J., Inglese, J., McGuire, J. J., Nimec, Z., and Coward, J. K. (1985b). g-Fluoro-
methotrexate: Synthesis and biological activity of a potent inhibitor of dihydrofolate
reductase with greatly diminished ability to form poly-g-glutamates. Proc. Natl. Acad. Sci.
USA 82, 25982602.
Galivan, J., Ryan, T. J., Chave, K., Rhee, M., Yao, R., and Yin, D. (2000). Glutamyl
hydrolase: Pharmacological role and enzymatic characterization. Pharmacol. Ther. 85,
207215.
Gelb, M. H., Svaren, J. P., and Abeles, R. H. (1985). Fluoro ketone inhibitors of hydrolytic
enzymes. Biochemistry 24, 18131817.
George, S., Cichowicz, D. J., and Shane, B. (1987). Mammalian folylpoly-g-glutamate
synthetase. 3. Specificity for folate analogues. Biochemistry 26, 522529.
Hart, B. P. (1995). The synthesis and biological evaluation of fluorinated glutamates, folates,
and antifolates. Ph.D. Dissertation. The University of Michigan, Ann Arbor, MI.
368 James K. Coward and John J. McGuire
Hart, B. P., and Coward, J. K. (1993). The synthesis of DL-3,3-difluoroglutamic acid from a
3-oxoprolinol derivative. Tetrahedron Lett. 34, 49174920.
Hart, B. P., Haile, W. H., Licato, N. J., Bolanowska, W. E., McGuire, J. J., and
Coward, J. K. (1996). Synthesis and biological activity of folic acid and methotrexate
analogues containing L-threo(2S,4S)-4-fluoroglutamic acid and DL-3,3-difluoroglutamic
Acid. J. Med. Chem. 39, 5665.
Hudlicky, M. (1993). Stereospecific syntheses of all four stereoisomers of 4-fluoroglutamic
acid. J. Fluorine Chem. 60, 193210.
Jencks, W. P., and Regenstein, J. (1976). In CRC Handbook of Biochemistry & Molecular
Biology (G. D. Fasman, ed.), pp. 305351. CRC Press, Inc, Cleveland, OH.
Kokuryo, Y., Kawata, K., Nakatani, T., Kugimiya, A., Tamura, Y., Kawada, K.,
Matsumoto, M., Suzuki, R., Kuwabara, K., Hori, Y., and Ohtani, M. (1997). Synthesis
and evaluation of novel fluorinated methotrexate derivatives for application to rheuma-
toid arthritis treatment. J. Med. Chem. 40, 32803291.
Konas, D. W., and Coward, J. K. (1999). Synthesis of L-4,4-difluoroglutamic acid via
electrophilic difluorination of a lactam. Org. Lett. 1, 21052107.
Konas, D. W., and Coward, J. K. (2001). Electrophilic fluorination of pyroglutamic acid
derivatives: Application of substrate-dependent reactivity and diastereoselectivity to the
synthesis of optically active 4-fluoroglutamic acids. J. Org. Chem. 66, 88318842.
Konas, D. W., Pankuch, J. J., and Coward, J. K. (2002). The synthesis of (2S)-4,4-
difluoroglutamyl g-peptides based on Garners aldehyde and fluoro-Reformatsky chem-
istry. Synthesis-Stuttgart 26162626.
Leroux, F. (2004). Atropisomerism, biphenyls, and fluorine: A comparison of rotational
barriers and twist angles. ChemBioChem. 5, 644649.
Li, H., Ryan, T. J., Chave, K. J., and Van Roey, P. (2002). Three-dimensional structure of
human g-glutamyl hydrolase: A class I glutamine amidohydrolase adapted for a complex
substrate. J. Biol. Chem. 277, 2452224529.
Licato, N. J., Coward, J. K., Nimec, Z., Galivan, J., Bolanowska, W. E., and McGuire, J. J.
(1990). Synthesis of N-[N-(4-deoxy-4-amino-10-methylpteroyl)-4-fluoroglutamyl]-g-
glutamate, an unusual substrate for folylpoly-g-glutamate synthetase and g-glutamyl
hydrolase. J. Med. Chem. 33, 10221027.
Liederer, B. M., and Borchardt, R. T. (2006). Enzymes involved in the bioconversion of
ester-based prodrugs. J. Pharm. Sci. 95, 11771195.
Lienhard, G. E., and Secemski, I. I. (1973). P
1
, P
5
, Di(adenosine-5
0
)pentaphosphate,
a potent multisubstrate inhibitor of adenylate kinase. J. Biol. Chem. 248, 11211123.
Liu, X., Hu, E., Tian, X., Mazur, A., and Ebetino, F. H. (2002). Enantioselective synthesis
of phosphinyl peptidomimetics via an asymmetric Michael reaction of phosphinic acids
with acrylate derivatives. J. Organomet. Chem. 646, 212222.
Longo, G. S. A., Gorlick, R., Tong, W. P., Lin, S. L., Steinherz, P., and Bertino, J. R.
(1997). g-Glutamyl hydrolase and folylpolyglutamate synthetase activities predict poly-
glutamylation of methotrexate in acute leukemias. Oncol. Res. 9, 259263.
Malachowski, W. P., and Coward, J. K. (1994a). The chemistry of phosphapeptides:
Investigations on the synthesis of phosphonamidate, phosphonate, and phosphinate
analogues of glutamyl-g-glutamate. J. Org. Chem. 59, 76257634.
Malachowski, W. P., and Coward, J. K. (1994b). The chemistry of phosphapeptides:
Formation of functionalized phosphonochloridates under mild conditions and their
reaction with alcohols and amines. J. Org. Chem. 59, 76167624.
Manne, V., Yan, N., Carboni, J. M., Tuomari, A. V., Ricca, C. S., Brown, J. G.,
Andahazy, M. L., Schmidt, R. J., Patel, D., Zahler, R., Weinmann, R., Der, C. J.,
et al. (1995). Bisubstrate inhibitors of farnesyltransferase: A novel class of specific inhibi-
tors of ras transformed cells. Oncogene 10, 17631779.
FPGS and GH Inhibitors 369
Matherly, L. H., and Goldman, D. (2003). Membrane transport of folates. Vitam. Horm. 66,
403456.
Mathieu, M., Debousker, G., Vincent, S., Viviani, F., Bamas-Jacques, N., and Mikol, V.
(2005). Escherichia coli FolC structure reveals an unexpected dihydrofolate binding site
providing an attractive target for anti-microbial therapy. J. Biol. Chem. 280, 1891618922.
McBurney, M. W., and Whitmore, G. F. (1974). Isolation and biochemical characterization
of folate deficient mutants of Chinese hamster cells. Cell 2, 173182.
McClure, W. R., and Chow, Y. (1980). The kinetics and processivity of nucleic acid
polymerases. Methods Enzymol. 64, 277.
McDermott, A. E., Creuzet, F., Griffin, R. G., Zawadzke, L. E., Ye, Q. Z., and
Walsh, C. T. (1990). Rotational resonance determination of the structure of an
enzyme-inhibitor complex: Phosphorylation of an (aminoalkyl)phosphinate inhibitor
of D-alanyl-D-alanine ligase by ATP. Biochemistry 29, 57675775.
McGuire, J. J., and Coward, J. K. (1984). Pteroylpolyglutamates: Biosynthesis, degradation,
and function. In Folates and Pterins (R. L. Blakley and S. J. Benkovic, eds.), Vol. 1,
pp. 135190. John Wiley & Sons, New York.
McGuire, J. J., and Coward, J. K. (1985). DL-threo-4-Fluoroglutamic acid: A chain terminat-
ing inhibitor of folylpolyglutamate synthetase. J. Biol. Chem. 260, 67476754.
McGuire, J. J., and Coward, J. K. (2003). Folylpolyglutamate synthetase as a target for
therapeutic intervention. Drugs Future 28, 967974.
McGuire, J. J., Hsieh, P., Coward, J. K., and Bertino, J. R. (1980). Enzymatic synthesis of
folylpolyglutamates. Characterization of the reaction and its products. J. Biol. Chem. 255,
57765788.
McGuire, J. J., Hsieh, P., Coward, J. K., and Bertino, J. R. (1983). In vitro methotrexate
polyglutamate synthesis by rat liver folylpolyglutamate synthetase and its inhibition by
bromosulfophthalein. In Folyl and Antifolyl Polyglutamates (I. D. Goldman,
B. A. Chabner, and J. R. Bertino, eds.), Vol. 163, pp. 199214. Plenum Press, New York.
McGuire, J. J., Hsieh, P., Franco, C. T., and Piper, J. R. (1986). Folylpolyglutamate
synthetase inhibition and cytotoxic effects of methotrexate analogs containing 2,o-
diaminoalkanoic acids. Biochem. Pharmacol. 35, 26072613.
McGuire, J. J., Graber, M., Licato, N., Vincenz, C., Coward, J. K., Nimec, Z., and
Galivan, J. (1989). Biochemical and growth inhibitory effects of the erythro and threo
isomers of g-fluoromethotrexate, a methotrexate analogue defective in polyglutamyla-
tion. Cancer Res. 49, 45174525.
McGuire, J. J., Haile, W. H., Bey, P., and Coward, J. K. (1990). DL-3,3-Difluoroglutamate:
An enhancer of folylpolyglutamate elongation. J. Biol. Chem. 265, 1407314079.
McGuire, J. J., Bolanowska, W. E., Coward, J. K., Sherwood, R. F., Russell, C. A., and
Felschow, D. M. (1991). Biochemical and biological properties of methotrexate analogs
containing D-glutamic acid or D-erythro,threo-4-fluoroglutamic acid. Biochem. Pharmacol.
42, 24002403.
McGuire, J. J., Hart, B. P., Haile, W. H., Rhee, M., Galivan, J., and Coward, J. K. (1995).
DL-b,b-difluoroglutamic acid mediates position-dependent enhancement or termination
of pteroylpoly (g-glutamate) synthesis catalyzed by folylpolyglutamate synthetase. Arch.
Biochem. Biophys. 321, 319328.
McGuire, J. J., Hart, B. P., Haile, W. H., Magee, K. J., Rhee, M., Bolanowska, W. E.,
Russell, C., Galivan, J., Paul, B., and Coward, J. K. (1996a). Biological properties of
fluoroglutamate-containing analogs of folates and methotrexate with altered capacities to
form poly (g-glutamate) metabolites. Biochem. Pharmacol. 52, 12951303.
McGuire, J. J., Tsukamoto, T., Hart, B. P., Coward, J. K., Kalman, T. I., and Galivan, J.
(1996b). Exploitation of folate and antifolate polyglutamylation to achieve selective
anticancer chemotherapy. Invest. New Drugs 14, 317323.
370 James K. Coward and John J. McGuire
McGuire, J. J., Haile, W. H., Valiaeva, N., Bartley, D., Guo, J., and Coward, J. K. (2003).
Potent inhibition of human folylpolyglutamate synthetase by a phosphinic acid mimic of
the tetrahedral reaction intermediate. Biochem. Pharmacol. 65, 315318.
Nair, S. A., Lee, B., and Hangauer, D. (1995). Synthesis of orthogonally protected
L-homocysteine and L-2-amino-4-phosphonobutanoic acid from L-homoserine.
Synthesis 7, 810814.
Pankuch, J. J. (2004). New fluorogenic peptide substrates for studying the mechanism of
g-glutamyl hydrolase. Ph.D. Dissertation. University of Michigan, Ann Arbor, MI.
Patel, D. V., Gordon, E. R., Schmidt, R. J., Weller, H. N., Young, M. G., Zahler, R.,
Barbacid, M., Carboni, J. M., Gullo-Brown, J. L., Hunihan, L., Ricca, C., Robinson, S.,
et al. (1995). Phosphinyl acid-based bisubstrate analog inhibitors of ras farnesyl protein
transferase. J. Med. Chem. 38, 435442.
Pizzorno, G., Mini, E., Coronnello, M., McGuire, J. J., Moroson, B. A., Cashmore, A. R.,
Dreyer, R. N., Lin, J. T., Mazei, T., Periti, P., and Bertino, J. R. (1988). Impaired
polyglutamylation of methotrexate as a cause of resistance in CCRF-CEM cells after
short-term, high-dose treatment with this drug. Cancer Res. 48, 21492155.
Qasba, P. K., Ramakrishnan, B., and Boeggeman, E. (2005). Substrate-induced conforma-
tional changes in glycosyltransferases. Trends Biochem. Sci. 30, 5362.
Rosowsky, A. (1999). PT523 and other aminopterin analogs with a hemiphthaloyl-L-
ornithine side chain: Exceptionally tight-binding inhibitors of dihydrofolate reductase
which are transported by the reduced folate carrier but cannot form polyglutamates. Curr.
Med. Chem. 6, 329352.
Rosowsky, A., Freisheim, J. H., Moran, R. G., Solan, V. C., Bader, H., Wright, J. E., and
Radike-Smith, M. (1986). Methotrexate analogues. 26. Inhibition of dihydrofolate
reductase and folylpolyglutamate synthetase activity and in vitro tumor cell growth by
methotrexate and aminopterin analogues containing a basic amino acid side chain. J. Med.
Chem. 29, 655660.
Rots, M. G., Pieters, R., Peters, G. J., Noordhuis, P., van Zantwijk, C. H., Kaspers, G. J. L.,
Hahlen, K., Creutzig, U., Veerman, A. J. P., and Jansen, G. (1999). Role of folylpo-
lyglutamate synthetase and folylpolyglutamate hydrolase in methotrexate accumulation
and polyglutamylation in childhood leukemia. Blood 93, 16771683.
Salcedo, E., Cortese, J. F., Plowe, C. V., Sims, P. F. G., and Hyde, J. E. (2001).
A bifunctional dihydrofolate synthetase-folylpolyglutamate synthetase in Plasmodium
falciparum identified by functional complementation in yeast and bacteria. Mol. Biochem.
Parasitol. 112, 239252.
Schirch, V., and Strong, W. B. (1989). Interaction of folylpolyglutamates with enzymes in
one-carbon metabolism. Arch. Biochem. Biophys. 269, 371380.
Schlosser, M. (1998). Parametrization of substituents: Effects of fluorine and other heteroa-
toms on OH, NH, and CH acidities. Angew. Chem. Int. Ed. 110, 14961513.
Schneider, E., and Ryan, T. J. (2006). g-Glutamyl hydrolase and drug resistance. Clinica
Chimica Acta 374, 2532.
Shane, B. (1989). Folylpolyglutamate synthesis and role in the regulation of one-carbon
metabolism. Vitam. Horm. 45, 263335.
Stokstad, E. L. R., and Koch, J. (1967). Folic acid metabolism. Physiol. Rev. 47, 83116.
Sun, X., Bognar, A. L., Baker, E. N., and Smith, C. A. (1998). Structural homologies with
ATP- and folate-binding enzymes in the crystal structure of folylpolyglutamate synthe-
tase. Proc. Natl. Acad. Sci. USA 95, 66476652.
Sun, X., Cross, J. A., Bognar, A. L., Baker, E. N., and Smith, C. A. (2001). Folate-binding
triggers the activation of folylpolyglutamate synthetase. J. Mol. Biol. 310, 10671078.
Suzuki, A., Mae, M., Amii, H., and Uneyama, K. (2004). Catalytic route to the synthesis of
optically active b,b-difluoroglutamic acid and b,b-difluoroproline derivatives. J. Org.
Chem. 69, 51325134.
FPGS and GH Inhibitors 371
Tang, K.-C., and Coward, J. K. (1983). Synthesis of b-keto phosphonate analogues of
biologically important acyl phosphates: N-(2-Amino-10-methylpteroyl)-5-amino-2-
oxopentane phosphonic acid. J. Org. Chem. 48, 50015006.
Taylor, R. T., and Hanna, M. L. (1977). Folate-dependent enzymes in cultured Chinese
hamster cells: Folylpolyglutamate synthetase and its absence in mutants auxotrophic for
glycine adenosine thymidine. Arch. Biochem. Biophys. 181, 331344.
Timperley, C. M., and White, W. E. (2003). The steric and electronic effect of aliphatic
fluoroalklyl groups. J. Fluorine Chem. 123, 6570.
Tomsho, J. W., Moran, R. G., and Coward, J. K. (2008). Concentration-dependent
processivity of multiple glutamate ligations catalyzed by folylpoly-g-glutamate synthetase.
Biochemistry, in press.
Tomsho, J. W., McGuire, J. J., and Coward, J. K. (2005). Synthesis of (6R)- and (6S)-5,
10-dideazatetrahydrofolate oligo-g-glutamates: Kinetics of multiple glutamate ligations
catalyzed by folylpoly-g-glutamate synthetase. Org. Biomol. Chem. 3, 33883398.
Tsukamoto, T., and Coward, J. K. (1996). Facile synthesis of DL-4,4-difluoroornithine,
DL-4,4-difluoroglutamine, and g-DL-4,4-difluoroglutamyl-containing peptides: Regios-
pecific addition of nucleophiles to N-Cbz-di-tert-butyl-DL-4,4-difluoroglutamate. J. Org.
Chem. 61, 24972500.
Tsukamoto, T., Coward, J. K., and McGuire, J. J. (1996a). Fluoroamino acid-containing
analogues of folic acid and methotrexate. In Biomedical Frontiers of Fluorine Chemistry
(I. Ojima, J. R. McCarthy, and J. T. Welch, eds.), pp. 118128. American Chemical
Society, Washington, DC.
Tsukamoto, T., Haile, W. H., McGuire, J. J., and Coward, J. K. (1996b). Synthesis and
biological evaluation of N
a
-(4-amino-4-deoxy-10-methylpteroyl)-DL-4,4-difluoroor-
nithine. J. Med. Chem. 39, 25362540.
Tsukamoto, T., Kitazume, T., McGuire, J. J., and Coward, J. K. (1996c). Synthesis and
biological evaluation of DL-4,4-difluoroglutamic acid and DL-g, g-difluoromethotrexate.
J. Med. Chem. 39, 6672.
Tsukamoto, T., Haile, W. H., McGuire, J. J., and Coward, J. K. (1998). Mechanism-based
inhibition of human folylpolyglutamate synthetase: Design, synthesis, and biochemical
characterization of a phosphapeptide mimic of the tetrahedral intermediate. Arch. Bio-
chem. Biophys. 355, 109118.
Tsukamoto, T., Flanary, J. M., Rojas, C., Slusher, B. S., Valiaeva, N., and Coward, J. K.
(2002). Phosphonate and phosphinate analogues of N-acylated g-glutamylglutamate:
Potent inhibitors of glutamate carboxypeptidase II. Bioorg. Med. Chem. Lett. 12,
21892192.
Tsushima, T., Kawada, K., Ishihara, S., Uchida, N., Shiratori, O., Higaki, J., and Hirata, M.
(1988). Fluorine-containing amino acids and their derivatives. 7. Synthesis and antitumor
activity of a- and g-substituted methotrexate analogs. Tetrahedron 44, 53755387.
Unkeless, J. C., and Goldman, P. (1971). The diastereomers of g-fluoroglutamate: Comple-
mentary structural analogues. Mol. Pharmacol. 7, 293300.
Valiaeva, N., Bartley, D., Konno, T., and Coward, J. K. (2001). Phosphinic acid pseudopep-
tides analogous to glutamyl-g-glutamate: Synthesis and coupling to pteroyl azides leads to
potent inhibitors of folylpoly-g-glutamate synthetase. J. Org. Chem. 66, 51465154.
Vidal-Cros, A., Gaudry, M., and Marquet, A. (1985). L-threo- and L-erythro-3-fluoroglutamic
acids. Synthesis by fluorodehydroxylation and enzymatic resolution. J. Org. Chem. 50,
31633167.
Vidal-Cros, A., Gaudry, M., and Marquet, A. (1989). Facile synthesis of optically pure
(2R,3R)- and (2R,3S)-3-fluoroglutamic acids using glutamate dehydrogenase. J. Org.
Chem. 54, 498500.
372 James K. Coward and John J. McGuire
Watanabe, K., Toh, Y., Suto, K., Shimizu, Y., Oka, N., Wada, T., and Tomita, K. (2007).
Protein-based peptide bond formation by aminoacyl-tRNA protein transferases. Nature
449, 867.
Wolfenden, R. (1972). Analog approaches to the structure of the transition-state in enzyme
reactions. Acct. Chem. Res. 5, 1018.
Wolfenden, R. (2003). Thermodynamic and extrathermodynamic requirements of enzyme
catalysis. Biophys. Chem. 105, 559572.
Yang, Y., and Coward, J. K. (2007). Synthesis of p-aminophenyl aryl H-phosphinic acids
and esters via cross-coupling reactions: Elaboration to phosphinic acid pseudopeptide
analogs of pteroyl glutamic acid and related antifolates. J. Org. Chem. 72, 57485758.
Yount, R. G. (1975). ATP analogs. Adv. Enzymol. Relat. Areas Mol. Biol. 43, 156.
FPGS and GH Inhibitors 373
C H A P T E R T H I R T E E N
Methylenetetrahydrofolate
Reductase, Common Polymorphisms,
and Relation to Disease
Philip Thomas* and Michael Fenech*
Contents
I. Introduction 376
A. Methylenetetrahydrofolate reductase 377
B. MTHFR and disease 381
C. MTHFR and pregnancy outcomes 383
D. MTHFRenvironment interactions 384
E. MTHFRother gene interactions 385
II. Conclusions 386
References 386
Abstract
Folate plays a key role in maintaining genomic stability and providing methyl
groups for the formation of dTMP from dUMP which is required for DNA
synthesis and repair and for the maintenance of methylation patterns involving
cytosine or specific sites such as CpG islands. Under conditions of low folate,
dUMP accumulates producing DNA strand breaks and micronucleus formation
as a result of excessive uracil incorporation into DNA in place of thymine.
Methylenetetrahydrofolate reductase (MTHFR) is an important folate metabo-
lizing enzyme that catalyzes the irreversible conversion of 5,10-methylenetre-
trahydrofolate, which is the methyl donor for the conversion of dUMP to dTMP,
into 5-methyltetrahydrofolate, which is the methyl donor for remethylation of
homocysteine to methionine. Certain common polymorphisms within the
MTHFR gene (C677T, A1298C) result in reduced enzymatic activity and have
been associated with reduced risk for a variety of cancers such as acute
lymphocytic leukemia, lung and colorectal cancer. These common polymorph-
isms are also associated with hyperhomocysteinemia that has been reported to
be an increased risk factor for neural tube defects and cardiovascular disease.
Vitamins and Hormones, Volume 79
#
2008 Elsevier Inc.
ISSN 0083-6729, DOI: 10.1016/S0083-6729(08)00413-5 All rights reserved.
* CSIRO Human Nutrition, Adelaide BC, Adelaide, South Australia 5000
375
In this chapter, we consider the role that MTHFR plays in relation to folate
metabolism and the possible contribution made in relation to certain important
clinical outcomes. 2008 Elsevier Inc.
I. Introduction
Folate (vitamin B
9
) is an essential B vitamin that is crucial to the
prevention of genomic instability and hypomethylation of DNA (Choi
and Mason, 2000, 2002). Folate is required for the synthesis of deoxythy-
midine monophosphate (dTMP) from deoxyuridine monophosphate
(dUMP), which is essential for DNA synthesis and repair (Fig. 13.1).
Under conditions of folate deficiency, dUMP accumulates resulting in
excessive uracil incorporation into DNA leading to single- and double-
strand DNA breaks, chromosome breakage, and ultimately micronucleus
(MN) formation (Blount and Ames, 1995; Blount et al., 1997; Fenech,
2001). Folate and vitamin B
12
are required in the synthesis of methionine
through the remethylation of homocysteine (Hcy) that ultimately leads to
the synthesis of S-adenosylmethionine (SAM). SAM plays an important role
as a methyl donor required for the maintenance of genomic methylation
patterns that determine gene expression, DNA conformation, and is
required for the synthesis of myelin, neurotransmitters, and membrane
phospholipids (Calvaresi and Bryan, 2001; Zingg and Jones, 1997). Folate
deficiency reduces SAM levels resulting in lower DNA cytosine methyla-
tion and elevated levels of Hcy. Additionally, folate deficiency may lead to
demethylation of centromeric DNA repeat sequences and centromere
dysfunction leading to abnormal chromosome distribution during nuclear
DNA synthesis
and repair
B
6
MTHFR
(B
2
)
Cystathione
Homocysteine
B
6
MTRR
5-methyl THF
dTMP
B
12
Cob (III)
B
12
Cob (II)
B
12
Cob (I)
Folic acid
SAM
Dietary folate
CH
3
DNA
methylation
Gene expression
Cell proliferation and
protein synthesis
THF
5,10-methyl
THF
DHF
5,10-methylene
THF
dUMP
MTR
Methionine
Figure 13.1 Metabolism of folic acid. Adapted from Wagner (1995). SAM: S-adenosyl
methionine, MTRR: methionine synthase reductase, MTR, methionine synthase;
THF, tetrahyrdofolate; DHF, dihydrofolate; MTHFR, methylene tetrahydrofolate
reductase; dUMP, deoxyuridine monophosphate; dTMP, deoxythymidine monopho-
sphate; Cob(I), reduced form of vitamin B
12
; Cob(III), oxidized form of vitamin B
12
.
376 Philip Thomas and Michael Fenech
division, resulting in elevated rates of aneuploidy, altered gene dosage and
increased cancer risk (Beetstra et al., 2005; Duesberg et al., 1999; Rasnick
and Duesberg, 1999; Wang et al., 2004).
A. Methylenetetrahydrofolate reductase
One of the intriguing aspects of the relationship between folate status
and cancer risk is the potential modifying effect of polymorphisms in key
folate metabolizingenzymes. Methylenetetrahydrofolate reductase (MTHFR)
is a pivitol enzyme within the folate methionine pathway which can influ-
ence both the bioavailability of folate for dTMP synthesis and maintain
methylation patterns at CpG islands known to regulate gene expression.
MTHFR catalyzes the reduction of 5,10-methylenetretrahydrofolate into
5-methyltetrahydrofolate, which is the major circulating form of folate, and
acts as a methyl donor in the remethylation of Hcy to methionine (Crott et al.,
2001; Goyette et al., 1998, 1994; Rozen, 1997; Sibani et al., 2003). The
MTHFR gene was cloned in 1998 and found to be 20.3 kb long, consisting
of 11 exons ranging in size from 102 to 432 bp (Frosst et al., 1995; Goyette
et al., 1998). The major gene product is a catalytically active 77 kDa protein
consisting of 656 amino acids, which has been shown to map to the short
arm of chromosome 1 at 1p36.3 (Goyette et al., 1994).
MTHFR enzymatic activity can be affected in a number of ways. First,
polymorphisms within the gene sequence could alter the affinity of the
enzyme for either substrate or its cofactor flavin adenine dinucleotide (FAD
or vitamin B
2
). Second, high concentrations of methionine or SAM are
inhibitory to MTHFRactivity, and finally, insufficient levels of FAD cofactor
may lead to reduced enzymatic activity (Hustad et al., 2000; Kimura et al.,
2004; Rivlin, 1996). Under conditions of reduced MTHFR activity, 5,10-
methylenetetrahydrofolate concentration increases with a resultant subsequent
lowering of 5-methyltetrahydrofolate concentration. This shift in balance
favors dTTP synthesis over CpG island methylation, a reduction in the
number of chromosome breaks by minimizing uracil incorporation, an
increase in DNA hypomethylation that could favor chromosome loss, and a
resultant increase in the concentration of plasma Hcy (Blount and Ames, 1995;
Blount et al., 1997; Castro et al., 2004; Kimura et al., 2004; Stern et al., 2000).
Many polymorphisms within the MTHFR gene have been reported
within the literature; however, very few have been conclusively studied in
relation to disease and population dynamics. The two most widely studied
of these are the common C677T and A1298C polymorphisms.
1. MTHFR C677T polymorphism
A common genetic variant of MTHFR involves a cytosine (C) to thymine
(T) transition at position 677 within exon 4 of the gene, resulting in an
alanine to valine substitution and reduced enzyme activity. This C677T
MTHFR, Common Polymorphisms, and Relation to Disease 377
transition is a mutation that causes thermolability as the activity of the
encoded enzyme is reduced at 37
C (Botto, 2000; Kang et al., 1991).
Enzyme activity of the CT heterozygote and the TT homozygote is
reduced by 35% and 70%, respectively, when compared to the normal
CC genotype ( James et al., 1999). Homozygosity for the T allele is asso-
ciated with reduced enzyme activity resulting in mild to moderately
elevated Hcy levels (Frosst et al., 1995).
The population frequency of the C677T allele shows significant differ-
ences based upon geographical location and ethnic background. Wilcken
et al. (2003) studied the C677T polymorphism in more than 7000 newborns
from 16 areas within Europe, Asia, the Americas, the Middle East, and
Australia. The TT homozygosity was particularly common in northern
China (20%), southern Italy (26%), and Mexico (32%). There was also
some evidence for changes in geographic gradients in Europe (north to
south increase) and China (north to south decrease). The TT genotype
frequency was low among new born individuals of African ancestry, inter-
mediate among newborns of European origin, and high among newborns of
American Hispanic ancestry. Areas at the extremes of the frequency distri-
bution showed deviations from Hardy-Weinberg expectations (Helsinki,
southern Italy, and southern China). The findings suggested the existence of
selective pressures leading to the marked variation. It has been shown from
other studies that C677T homozygosity in white Europeans ranged from
7.2% in Germany to 22% in northern Italy (Cattaneo et al., 1997; Koch
et al., 1998). In Japanese populations, the homozygous frequency was
shown to range between 10.2% and 12.2% whereas in black Africans, the
homozygosity was not reported but the allelic frequency was 7% (Arinami
et al., 1997; Morita et al., 1998; Schneider et al., 1998). This suggests that the
allele is less common within the black African population. It is plausible that
dietary folate may exert a selective pressure if high maternal folate deter-
mines the survival of C677T TT homozygotes in utero; preliminary data
suggests that this may be possible but evidence is as yet inconclusive (Lucock
et al., 2003; Munoz-Moran et al., 1998).
2. MTHFR A1298C polymorphism
A second polymorphism in the MTHFR gene involves an A to C transition
at position 1298 within exon 7 that results in a change from a glutamate to
an alanine residue. This mutation alters an mboII recognition site and has an
allelic frequency of 0.33 (van der Put et al., 1998). The activity of the
enzyme is decreased but not to the same extent as the C677T allele
(Weisberg et al., 1998a). It has been shown that neither the homozygous
nor the heterozygous state of A1298C leads to an elevation in plasma Hcy or
lower plasma folate concentration, which is evident for the homozygous
state for C677T (Lievers et al., 2001). However, compound heterozygosity
for both the C677T and the A1298C is associated with reduced MTHFR
378 Philip Thomas and Michael Fenech
enzyme activity, higher plasma Hcy, and lower plasma folate concentrations
(Botto et al., 2003; Friedman et al., 1999; Weisberg et al., 1998b).
The combined association of these two alleles produces a similar biochemi-
cal profile to those individuals who are homozygote for the T allele of
C677T. The population frequency for A1298C homozygosity is not as well
documented as for the C677T allele and is thought to have a prevalence of
about 10% (van der Put et al., 1998; Weisberg et al., 1998a). To date, over
30 other mutations have been identified as being associated with severe
MTHFR deficiency (Sibani et al., 2000, 2003). The relationship between
these rare polymorphisms and population dynamics and disease outcomes
has yet to be fully determined.
3. MTHFR and genomic instability
Damage to the genome could lead to altered gene dosage and gene expres-
sion as well as contribute to the risk of accelerated cellular death. Certain
genomic instability biomarkers have been found to be altered in individuals
possessing polymorphisms within the MTHFR gene. MN formation is a
biomarker of chromosome malsegregation and fragility and has been found
to be elevated in individuals carrying the variant C677T, TT genotype
compared with the wild-type or heterozygous genotype (Andreassi et al.,
2003; Botto et al., 2003; Kimura et al., 2004). Nucleoplasmic bridges
(a biomarker for chromosomal rearrangement) were also found to be sign-
ificantly increased in C677T, TT cells grown under low folate conditions
compared with C677C or C677T cells (Leopardi et al., 2006). Kimura et al.
also showed that nuclear buds (a biomarker for gene amplification) were
markedly higher under conditions of low folate and high riboflavin com-
pared with low folate and low riboflavin conditions, indicating a potential
genotoxic effect of elevated riboflavin concentrations under conditions of
low folate. Note that the level of nuclear buds was lower in C677T, TT cells
compared with C677C, or C677T cells. This suggests that nuclear bud
formation under low folate conditions may be exacerbated when riboflavin
concentration increases or in the presence of the MTHFR wild-type
genotype, both of which increase MTHFR activity (Kimura et al., 2004).
Abnormal folate and methyl metabolism have been shown to lead to
chromosome malsegregation and DNA hypomethylation (Fenech, 2001).
Hypomethylation of repeat sequences within the centromeric regions of
chromosomes may lead to faulty kinetochore assembly or despiralization
leading to the loss of chromosomes as micronuclei (Botto et al., 2003). It has
been shown that individuals who have the C677T, TT polymorphism are
associated with genomic DNA hypomethylation and have an increased risk
for having children with three copies of chromosome 21 (Down syndrome)
(Friso et al., 2002; Stern et al., 2000). James et al. (1999) found a 2.6-fold
higher risk in individuals with the C677T polymorphism of having a Down
syndrome child compared with wild-type individuals. It has also been
MTHFR, Common Polymorphisms, and Relation to Disease 379
reported that mothers of Down syndrome children have a higher frequency
of joint heterozygotic MTHFR polymorphisms (677 and 1298) compared
to those with chromosomally normal offspring (OR: 5.7) (Gregorio
Lorenzo Acacio, 2005).
The interactive relationships between folate, MTHFR genotype, and its
cofactor riboflavin are complex and have been shown to affect a number of
biomarkers of genomic instability. A flow diagram of these relationships and
potential outcomes are highlighted in Fig. 13.2.
The framework predicts that (1) as folate concentration increases, geno-
mic instability events arising from aneuploidy and breakage-fusion-bridge
(BFB) cycles are minimized, (2) genome hypomethylation and aneuploidy
are minimized during high MTHFR activity, (3) the risk of BFB cycles is
increased under high riboflavin and low folate conditions, (4) the risk of
genome hypomethylation and aneuploidy are maximized under both low
riboflavin and low folate conditions, and (5) reduced MTHFR activity may
decrease MN frequency caused by uracil incorporation and subsequent
chromosome breakage but may increase MN originating from chromosome
loss or gain resulting from changes in methylation patterns. These complex
Low
MTHFR
activity
High
MTHFR
activity
Low CpG
methylation *
High CpG
methylation *
Low uracil in DNA
High uracil in DNA
High
Low F
Low F
Low R
High R
G
E
N
O
T
Y
P
E
High MNi (Chromosome
Loss/Gain) frequency
increased
ANEUPLOIDY
Low MNi (Chromosome
loss/gain) frequency
decreased
ANEUPLOIDY
Low MNi (Chromosome breaks)
Low NPB (Chromosome
rearrangement)
Low NBUDs (Gene
amplification)
High MNi (Chromosome breaks)
High NPB (Chromosome
rearrangement)
High NBUDs (Gene
amplification)
REDUCED
CANCER
RISK?
INCREASED
CANCER
RISK?
INCREASED
CANCER
RISK?
Figure 13.2 Mechanistic framework explaining the interrelationship between
MTHFR genotype, riboflavin (R), and folic acid (F) with respect to (1) CpG methyla-
tion and uracil in DNA, (2) aneuploidy and micronuclei (MNi) originating from
chromosome loss events, (3) MNi (originating from acentric chromosome fragments),
nuclear buds (NBUDs), nucleoplasmic bridges (NPBs), and breakage-fusion-bridge
(BFB) cycles, (4) initiation of cancer caused by CpG hypomethylation and aneuploidy,
and (5) initiation of cancer caused by increased BFB cycles, MNi (originating from
acentric chromosome fragments), NBUDs, and NPBs. *For brevity, other carcinogenic
mechanisms induced by altered genome methylation, such as silencing of tumor
suppressor genes and/or activation of oncogenes, are not included in the diagram.
380 Philip Thomas and Michael Fenech
relationships may go some way to explain how the MTHFR C677T
polymorphism may influence the relative risk for certain important clinical
outcomes. The prevention of chromosome breakage and BFB cycles caused
by uracil incorporation into the DNA may be more relevant to conditions
such as leukemia or lymphoma, whereas prevention of CpG hypomethyla-
tion may be more relevant to cancers caused by DNA hypomethylation or
aneuploidy-driven developmental defects such as Down syndrome.
B. MTHFR and disease
1. Cancer
Low folate intake has been shown to be associated with increased cancer risk
(Giovannucci et al., 1993, 1995; Sellers et al., 2001; Zhang et al., 1999), raising
the possibility of a potential role for the C677T mutation in carcinogenesis or
cancer progression. The association between MTHFR and cancer suscepti-
bility has been examined in a number of different cancers. Individuals with
C677T homozygosity have been shown to have a 2.8-fold increased risk for
endometrial cancer, whereas patients with ovarian tumors have been shown to
have allelic deletions in the MTHFR gene (Esteller et al., 1997; Viel et al.,
1997). In contrast, various studies have shown that TT homozygotes have a
1.2- to 3.0-fold reduced risk for colorectal cancer and a 4.3-fold reduced risk
for acute lymphocytic leukemia, but the protective effect may be lost in
individuals who are folate deficient (Chen et al., 1996; Ma et al., 1997;
Skibola et al., 1999). Reduced MTHFR enzymatic activity may prove to be
protective by inhibiting hypermethylation that could lead to CpG island
silencing of certain tumor suppressor genes. For example, it has been shown
in patients with lung cancer that the TT allele is associated with the increased
expression of the tumor suppressor gene p16 (Kamiya et al., 1998).
It is also thought that the protective effect afforded by the polymorphism
toward conditions such as lung and colorectal cancers may be due to the
diversion of folate to purine and thymidine synthesis. This leads to the
increased availability of 5,10-methylenetetrahydrofolate and subsequent
methyl groups for the conversion of dUMP to dTMP, thereby reducing
uracil incorporation and subsequent DNA instability (Crott et al., 2001;
Kimura et al., 2004; Ueland et al., 2001).
A small number of studies have investigated the association between the
A1298C polymorphism and colorectal cancer risk (Chen et al., 2002; Keku
et al., 2002; Le Marchand, 2002). In all studies, a reduced risk was found to
be evident in CC individuals compared to AA subjects. Relative risks were
in the range of 0.60.8 but were not found to be statistically significant. It
has also been reported that the A1298C result was not due to a confounding
C677T effect (Chen et al., 2002) and that a greater protective effect
occurred in individuals who carried both 677T and 1298C alleles compared
to wild-type subjects (Le Marchand, 2002).
MTHFR, Common Polymorphisms, and Relation to Disease 381
2. Cardiovascular disease
Homozygosity for the C677T allele leads to a relative deficiency in the
remethylation process of Hcy into methionine resulting in mild-to-moder-
ate hyperhomocysteinemia, a condition recognized as an independent risk
factor for atherosclerosis (Clarke et al., 1991; Danesh and Lewington, 1998).
A number of studies have investigated whether an association between
MTHFR polymorphisms and cardiovascular disease exists. A meta-analysis
was performed in 1996 combining all studies where MTHFR genotyping
data was available for patients with coronary heart disease. It was shown that
individuals who were 677T homozygotes had a significant 19% higher risk
for coronary heart disease compared with the other MTHFR genotypes
(Blom, 1998). A more recent meta-analysis of over 72 studies in which
MTHFR genotypes were available in patients that had been diagnosed with
ischemic heart disease, deep vein thrombosis, or pulmonary embolism
showed a significantly higher risk in people with the MTHFR mutation
(Wald et al., 2002). Further meta-analysis investigation involving over 80
studies examining increased risk for coronary disease showed an odds ratio
of 1.14 (95% confidence intervals (CI): 1.051.24) for TT versus CC
genotype (Lewis et al., 2005). Although there is not a strong allelic associa-
tion in most studies, it may be that the resulting elevated Hcy may also
interact with other risk factors that may induce vascular events in those
individuals who have underlying conditions such as thrombophilia that may
predispose to cardiovascular disease (Refsum and Ueland, 1998; Ueland
et al., 2001).
3. Alzheimers disease
Alzheimers disease (AD) is a neurodegenerative disorder that is character-
ized clinically by cognitive decline, memory loss, visuospatial and language
impairment, and is the commonest form of dementia (Burns et al., 2002;
Kawas, 2003; Mattson, 2004; St George-Hyslop, 2000). Regions of the
brain that are involved in short-term memory and learning such as the
temporal and frontal lobes are impaired as a result of neuronal loss and
the breakdown of the neuronal synaptic connections (Mattson et al., 1998).
Many case control studies have shown that Alzheimers patients have
been found to be deficient in certain micronutrients such as folate, vitamin
B
12
, and have elevated levels of the sulfur-based amino acid Hcy (Aisen
et al., 2003; Shea and Rogers, 2002). These factors are associated with
increased MN formation and the alteration of methylation patterns that
could modify gene expression (Fenech and Crott, 2002; Scarpa et al., 2003;
Suzuki et al., 2002). It has been shown that adequate folate intake improves
global cognitive function (de Lau et al., 2007) and that increased Hcy levels
have a significant effect in the reduction of effective cognitive function
(Kim et al., 2007).
382 Philip Thomas and Michael Fenech
Hyperhomocysteinemia has been shown to be a strong independent risk
factor for AD in a number of epidemiological studies (Clarke et al., 1998;
McCaddon et al., 1998; Morris, 2003; Seshadri et al., 2002; Wang et al.,
2001). It appears that nervous tissue may be extremely sensitive to excessive
Hcy as it promotes excitotoxicity and damages neuronal DNA giving rise to
apoptosis (Kruman et al., 2002). Studies have also shown a strong correlation
between a reduction in hippocampal width, which is associated with short-
term memory loss and concentrations of plasma Hcy (Williams et al., 2002).
Recently MRI measurements have shown that an inverse relationship exists
between plasma Hcy and cortical and hippocampal volume (den Heijer
et al., 2003). The above findings have been interpreted as involving neuro-
nal damage within the hippocampal regions leading to memory loss, which
is characteristic of AD. Elevated Hcy has also been implicated as playing a
role in an iron dysregulation/oxidative stress cycle that is thought to be
central to the pathogenesis of the disease (Dwyer et al., 2004).
As AD is related to low folate, vitamin B
12
, and elevated Hcy levels,
investigators have looked toward genetic polymorphisms within the folate
methionine pathway (Fig. 13.1) to explain the effect of micronutrient
differences. In determining the risk factor of C677T in relation to AD, no
association between C677T and increased susceptibility to Alzheimers risk
was found (Brunelli et al., 2001). However, another study found that female
TT homozygotes have significant cognitive decline compared to wild type
and heterozygotes (Elkins et al., 2007). However, if combinations of poly-
morphisms within a gene are considered together, then sometimes effects
become apparent that are not always evident if those same polymorphisms
are considered in isolation. Wakutani et al. (2004) found that the MTHFR
677C-1298C-1793G haplotype to be protective in Japanese populations
against late onset AD.
C. MTHFR and pregnancy outcomes
1. Neural tube defects
Under low folate conditions, individuals possessing C677T polymorphisms
are predisposed to hyperhomocysteinemia. It has been shown that hyperho-
mocysteinemia is an increased risk factor for neural tube defects (NTD)
(Finnell et al., 1998; Steegers-Theunissen et al., 1994, 1995). Initial studies
reported an increased frequency of the C677Tallele in both affected mothers
and children inferring an increased risk for NTD. Data generated frommeta-
analysis studies indicate that TT individuals have a doubling in risk of having
an affected child (Botto, 2000). Interestingly, it has been observed that low
maternal blood folate and no periconceptional folate supplementation ele-
vate the risk associated with the T allele (Christensen et al., 1999; Shaw et al.,
1998). The combination of a TT genotype and low folate concentration
increases Hcy concentration, while impairing the formation of embryonic
MTHFR, Common Polymorphisms, and Relation to Disease 383
methionine which may result in abnormalities in myelin synthesis that
could contribute to NTD formation (Aubard et al., 2000).
Limited data is available regarding the association between A1298C and
the incidence of NTD. In one study, a significant association was found in a
subset of cases between the allele and an increased spina bifida risk but this
result could not be replicated (Botto, 2000; Trembath et al., 1999).
Compound heterozygosity for both the C677T and A1298C alleles may
increase spina bifida risk compared to the wild-type combinations. Two
separate studies have calculated odds ratios of 2.0 (95% CI: 0.95.1) and
2.8 (95% CI: 1.17.6) for allelic compound heterozygosity in relation to
increased risk for spina bifida (Botto, 2000; Trembath et al., 1999; van der
Put et al., 1998).
2. Preeclampsia
Hyperhomocysteinemia in conjunction with a TT genotype has been
associated with an increased risk for spontaneous abortion, placental abrup-
tion, and preeclampsia (Ray and Laskin, 1999). It is thought that the
elevated levels of Hcy produce placental endovascular damage as a result
of oxidative stress. It has been shown that 17% of women suffering with
severe preeclampsia were found to have hyperhomocysteinemia compared
with a 2% prevalence found in the general population (Dekker et al., 1995;
Leeda et al., 1998). Recent studies have not been able to associate a clear
relationship between MTHFR polymorphisms and placenta-mediated dis-
eases such as preeclampsia (Els et al., 2000; Laivuori et al., 2000; Thomas
Kaiser and Moses, 2000). However, an increased incidence of the TT allele
has been reported in women who have experienced abruption placentae,
interuterine fetal growth retardation, and still births compared to women
who have had normal pregnancies (Kupferminc et al., 1999). These pregnancy
outcomes are associated with elevated Hcy and emphasizes the importance of
maintaining adequate folate levels periconceptionally, especially in individuals
who possess MTHFR polymorphisms that may contribute to increased risk
for abnormal pregnancies.
D. MTHFRenvironment interactions
A number of studies have examined the interaction between nutritional
status and the C677T polymorphism in relation to clinical outcomes. Two
studies suggest that neural tube risk associated with C677T, TT, homozy-
gosity may be dependent on nutritional status. The first study showed a
13-fold increased risk for spina bifida in C677T, TT, homozygotes with a
red blood cell folate value in the lowest study quartile (Christensen et al.,
1999). The second found that maternal multivitamin use (containing folic
acid) reduced the risk for spina bifida in individuals with wild-type or
heterozygous alleles (OR 0.3) and in those with the homozygous TT
384 Philip Thomas and Michael Fenech
genotype (OR 0.2) (Shaw et al., 1998). Studies investigating colorectal
cancer showed that the inverse association with TT homozygosity was
greatest in those individuals with high folate or methionine intake (Chen
et al., 1996; Ma et al., 1997). These results suggest that dietary methyl supply
is particularly important among TTindividuals. When dietary methyl supply
is high, TT individuals may be at reduced risk of colorectal cancer as the
higher levels of 5,10-methylenetetrahydrofolate make purines and thymi-
dines available for the nucleotide pool during DNA synthesis and repair.
Alcohol consumption was found to remove the reduced risk associated
with TT individuals for colorectal cancer (Ma et al., 1997). In fact TT
individuals who consumed large volumes of alcohol were at even greater
risk than those without the Tallele whoconsumed similar volumes of alcohol
(Chen et al., 1996). This would suggest that when 5-methyltetrahydrofolate
is depleted by alcohol consumption, TT individuals may be less able to
compensate, leading to alterations in methylation patterns that may lead to
altered expression of oncogenes.
E. MTHFRother gene interactions
Other genes whose alleles have been studied in combination with MTHFR
alleles include transcobalamin, cystathionine-b-synthase, methionine syn-
thase reductase, and methionine synthase. Polymorphisms in the MTHFR
and transcobalamin genes (C776G) have been shown to influence Hcy
metabolism that may in turn increase the risk for spontaneous abortion. It
has been shown that embryos with a combined T677T and transcobalamin
C776G or G776G genotype have an odds ratio of 3.8 for spontaneous
abortion compared with embryos with only one of these genotypes
(Zetterberg et al., 2003). Cystathionine-b-synthase catalyzes the conversion
of Hcy to cystathionine, thus providing an alternative route for Hcy
metabolism. It has been shown that individuals who have polymorphisms
for both the C677T allele and the cystathionine-b-synthase 844ins68 allele
are at an increased risk for spina bifida (OR 5.2; 95% CI: 1.421.2). The
risk is higher than would be expected if each of the polymorphisms for
C677T (OR 2.1; 95% CI: 1.13.9) or 844ins68 (OR 0.8; 95% CI:
0.41.4) were considered individually suggesting a significant genegene
interaction (Ramsbottom et al., 1997). It has been shown that individuals
who are homozygous for polymorphisms within the methionine synthase
reductase gene (A66G) and MTHFR C677T appear to have an increased
risk for spina bifida (OR 4.1; 95% CI: 1.016.4) (Morrison et al., 1998).
Further evidence for potential genegene interactions involving MTHFR
was shown in individuals that possess both the C677T, T allele, together
with the methionine synthase A2756G, G allele. This combination was
found to reduce the risk of colorectal cancer and may afford a certain degree
of protection (Le Marchand, 2002).
MTHFR, Common Polymorphisms, and Relation to Disease 385
II. Conclusions
The MTHFR C677T polymorphism has been implicated in a number
of clinical diseases. Under conditions of low folate, the MTHFR C677T,
TT genotype is associated with increased risk for NTDs. Conversely when
folate levels are adequate, the TT genotype may reflect a selective advantage
for the reduction in risk for both lung and colorectal cancer. Individuals who
are TT homozygotes or compound heterozygotes for both the A1298C and
C677T alleles may have to pay more attention to adjusting their folate and
riboflavin intake in order to reduce genomic instability and the effect of
potential clinical outcomes associated with hyperhomocysteinemia.
REFERENCES
Aisen, P. S., Egelko, S., Andrews, H., Diaz-Arrastia, R., Weiner, M., DeCarli, C.,
Jagust, W., Miller, J. W., Green, R., Bell, K., and Sano, M. (2003). A pilot study of
vitamins to lower plasma homocysteine levels in Alzheimer disease. Am. J. Geriatr.
Psychiatry 11, 246249.
Andreassi, M. G., Botto, N., Cocci, F., Battaglia, D., Antonioli, E., Masetti, S., Manfredi, S.,
Colombo, M. G., Biagini, A., and Clerico, A. (2003). Methylenetetrahydrofolate reduc-
tase gene C677T polymorphism, homocysteine, vitamin B12, and DNA damage in
coronary artery disease. Hum. Genet. 112, 171177.
Arinami, T., Yamada, N., Yamakawa-Kobayashi, K., Hamaguchi, H., and Toru, M. (1997).
Methylenetetrahydrofolate reductase variant and schizophrenia/depression. Am. J. Med.
Genet. 74, 526528.
Aubard, Y., Darodes, N., and Cantaloube, M. (2000). Hyperhomocysteinemia and preg-
nancyreview of our present understanding and therapeutic implications. Eur. J. Obstet.
Gynecol. Reprod. Biol. 93, 157165.
Beetstra, S., Thomas, P., Salisbury, C., Turner, J., and Fenech, M. (2005). Folic acid
deficiency increases chromosomal instability, chromosome 21 aneuploidy and sensitivity
to radiation-induced micronuclei. Mutat. Res. 578, 317326.
Blom, H. J. (1998). Mutated 510 methylenetetrahydrofolate reductase and moderate
hyperhomocysteinemia. Eur. J. Pediatr. 157, S131S134.
Blount, B. C., and Ames, B. N. (1995). DNA damage in folate deficiency. Baillieres Clin.
Haematol. 8, 461478.
Blount, B. C., Mack, M. M., Wehr, C. M., MacGregor, J. T., Hiatt, R. A., Wang, G.,
Wickramasinghe, S. N., Everson, R. B., and Ames, B. N. (1997). Folate deficiency causes
uracil misincorporation into human DNA and chromosome breakage: Implications for
cancer and neuronal damage. Proc. Natl. Acad. Sci. USA 94, 32903295.
Botto, L. D., and Yanq, Q. (2000). 5,10-Methylenetetrahydrofolate reductase gene variants
and congenital anomalies: AHuGE review. Am. J. Epidemiol. 151, 862877.
Botto, N., Andreassi, M. G., Manfredi, S., Masetti, S., Cocci, F., Colombo, M. G.,
Storti, S., Rizza, A., and Biagini, A. (2003). Genetic polymorphisms in folate and
homocysteine metabolism as risk factors for DNA damage. Eur. J. Hum. Genet. 11,
671678.
386 Philip Thomas and Michael Fenech
Brunelli, T., Bagnoli, S., Giusti, B., Nacmias, B., Pepe, G., Sorbi, S., and Abbate, R. (2001).
The C677T methylenetetrahydrofolate reductase mutation is not associated with Alzhei-
mers disease. Neurosci. Lett. 315, 103105.
Burns, A., Byrne, E. J., and Maurer, K. (2002). Alzheimers disease. Lancet 360, 163165.
Calvaresi, E., and Bryan, J. (2001). B vitamins, cognition, and aging: A review. J. Gerontol. B
Psychol. Sci. Soc. Sci. 56, P327P339.
Castro, R., Rivera, I., Ravasco, P., Camilo, M. E., Jakobs, C., Blom, H. J., and de
Almeida, I. T. (2004). 5,10-methylenetetrahydrofolate reductase (MTHFR) 677C!T
and 1298A!C mutations are associated with DNA hypomethylation. J. Med. Genet. 41,
454458.
Cattaneo, M., Tsai, M. Y., Bucciarelli, P., Taioli, E., Zighetti, M. L., Bignell, M., and
Mannucci, P. M. (1997). A common mutation in the methylenetetrahydrofolate reduc-
tase gene (C677T) increases the risk for deep-vein thrombosis in patients with mutant
factor V (factor V:Q506). Arterioscler. Thromb. Vasc. Biol. 17, 16621666.
Chen, J., Giovannucci, E., Kelsey, K., Rimm, E. B., Stampfer, M. J., Colditz, G. A.,
Spiegelman, D., Willett, W. C., and Hunter, D. J. (1996). A methylenetetrahydrofolate
reductase polymorphism and the risk of colorectal cancer. Cancer Res. 56, 48624864.
Chen, J., Ma, J., Stampfer, M. J., Palomeque, C., Selhub, J., and Hunter, D. J. (2002).
Linkage disequilibrium between the 677C > T and 1298A > C polymorphisms in
human methylenetetrahydrofolate reductase gene and their contributions to risk of
colorectal cancer. Pharmacogenetics 12, 339342.
Choi, S. W., and Mason, J. B. (2000). Folate and carcinogenesis: An integrated scheme.
J. Nutr. 130, 129132.
Choi, S. W., and Mason, J. B. (2002). Folate status: Effects on pathways of colorectal
carcinogenesis. J. Nutr. 132, 2413S2418S.
Christensen, B., Arbour, L., Tran, P., Leclerc, D., Sabbaghian, N., Platt, R., Gilfix, B. M.,
Rosenblatt, D. S., Gravel, R. A., Forbes, P., and Rozen, R. (1999). Genetic polymorph-
isms in methylenetetrahydrofolate reductase and methionine synthase, folate levels in red
blood cells, and risk of neural tube defects. Am. J. Med. Genet. 84, 151157.
Clarke, R., Daly, L., Robinson, K., Naughten, E., Cahalane, S., Fowler, B., and Graham, I.
(1991). Hyperhomocysteinemia: An independent risk factor for vascular disease. N. Engl.
J. Med. 324, 11491155.
Clarke, R., Smith, A. D., Jobst, K. A., Refsum, H., Sutton, L., and Ueland, P. M. (1998).
Folate, vitamin B12, and serum total homocysteine levels in confirmed Alzheimer
disease. Arch. Neurol. 55, 14491455.
Crott, J. W., Mashiyama, S. T., Ames, B. N., and Fenech, M. F. (2001). Methylenetetrahy-
drofolate reductase C677T polymorphism does not alter folic acid deficiency-induced
uracil incorporation into primary human lymphocyte DNA in vitro. Carcinogenesis 22,
10191025.
Danesh, J., and Lewington, S. (1998). Plasma homocysteine and coronary heart disease:
Systematic review of published epidemiological studies. J. Cardiovasc. Risk 5, 229232.
de Lau, L. M. L., Refsum, H., Smith, A. D., Johnston, C., and Breteler, M. M. B. (2007).
Plasma folate concentration and cognitive performance: Rotterdam Scan Study. Am. J.
Clin. Nutr. 86, 728734.
Dekker, G. A., de Vries, J. I., Doelitzsch, P. M., Huijgens, P. C., von Blomberg, B. M.,
Jakobs, C., and van Geijn, H. P. (1995). Underlying disorders associated with severe
early-onset preeclampsia. Am. J. Obstet. Gynecol. 173, 10421048.
den Heijer, T., Vermeer, S. E., and Clarke, R. (2003). Homocysteine and brain atrophy on
MRI of non-demented elderly. Brain 126, 170175.
Duesberg, P., Rasnick, D., Li, R., Winters, L., Rausch, C., and Hehlmann, R. (1999). How
aneuploidy may cause cancer and genetic instability. Anticancer Res. 19, 48874906.
MTHFR, Common Polymorphisms, and Relation to Disease 387
Dwyer, B. E., Raina, A. K., Perry, G., and Smith, M. A. (2004). Homocysteine and
Alzheimers disease: A modifiable risk? Free Radic. Biol. Med. 36, 14711475.
Elkins, J. S., Johnston, S. C., Ziv, E., Kado, D., Cauley, J. A., and Yaffe, K. (2007).
Methylenetetrahydrofolate reductase C677T polymorphism and cognitive function in
older women. Am. J. Epidemiol. 166, 672678.
Els, F.v.d.M., Guus, E. A., Willianne, L. D. M. N., Nathalie, J. M. v. d. P., Sandra, G. H.,
Tom, K. A. B. E., and Henk, J. B. (2000). A common mutation in the 5,10-methyle-
netetrahydrofolate reductase gene as a new risk factor for placental vasculopathy. Am. J.
Obstet. Gynecol. 182, 12581263.
Esteller, M., Garcia, A., Martinez-Palones, J. M., Xercavins, J., and Reventos, J. (1997).
Germ line polymorphisms in cytochrome-P450 1A1 (C4887 CYP1A1) and methylene-
tetrahydrofolate reductase (MTHFR) genes and endometrial cancer susceptibility.
Carcinogenesis 18, 23072311.
Fenech, M. (2001). The role of folic acid and Vitamin B12 in genomic stability of human
cells. Mutat. Res. 475, 5767.
Fenech, M., and Crott, J. W. (2002). Micronuclei, nucleoplasmic bridges and nuclear buds
induced in folic acid deficient human lymphocytes-evidence for breakage-fusion-bridge
cycles in the cytokinesis-block micronucleus assay. Mutat. Res. 504, 131136.
Finnell, R. H., Greer, K. A., Barber, R. C., and Piedrahita, J. A. (1998). Neural tube and
craniofacial defects with special emphasis on folate pathway genes. Crit. Rev. Oral Biol.
Med. 9, 3853.
Friedman, G., Goldschmidt, N., Friedlander, Y., Ben-Yehuda, A., Selhub, J., Babaey, S.,
Mendel, M., Kidron, M., and Bar-On, H. (1999). A common mutation A1298C in
human methylenetetrahydrofolate reductase gene: Association with plasma total homo-
cysteine and folate concentrations. J. Nutr. 129, 16561661.
Friso, S., Choi, S. W., Girelli, D., Mason, J. B., Dolnikowski, G. G., Bagley, P. J.,
Olivieri, O., Jacques, P. F., Rosenberg, I. H., Corrocher, R., and Selhub, J. (2002).
A common mutation in the 5,10-methylenetetrahydrofolate reductase gene affects
genomic DNA methylation through an interaction with folate status. Proc. Natl. Acad.
Sci. USA 99, 56065611.
Frosst, P., Blom, H. J., Milos, R., Goyette, P., Sheppard, C. A., Matthews, R. G.,
Boers, G. J., den Heijer, M., Kluijtmans, L. A., van den Heuvel, L. P., and Rozen, R.
(1995). A candidate genetic risk factor for vascular disease: A common mutation in
methylenetetrahydrofolate reductase. Nat. Genet. 10, 111113.
Giovannucci, E., Rimm, E. B., Ascherio, A., Stampfer, M. J., Colditz, G. A., and
Willett, W. C. (1995). Alcohol, low-methionine-low-folate diets, and risk of colon
cancer in men. J. Natl. Cancer Inst. 87, 265273.
Giovannucci, E., Stampfer, M. J., Colditz, G. A., Rimm, E. B., Trichopoulos, D.,
Rosner, B. A., Speizer, F. E., and Willett, W. C. (1993). Folate, methionine, and alcohol
intake and risk of colorectal adenoma. J. Natl. Cancer Inst. 85, 875883.
Goyette, P., Pai, A., Milos, R., Frosst, P., Tran, P., Chen, Z., Chan, M., and Rozen, R.
(1998). Gene structure of human and mouse methylenetetrahydrofolate reductase
(MTHFR). Mamm. Genome 9, 652656.
Goyette, P., Sumner, J. S., Milos, R., Duncan, A. M., Rosenblatt, D. S., Matthews, R. G.,
and Rozen, R. (1994). Human methylenetetrahydrofolate reductase: Isolation of cDNA,
mapping and mutation identification. Nat. Genet. 7, 195200.
Gregorio Lorenzo Acacio, R. B., Bertuzzo, C. S., Couto, E. C., Annichino-
Bizzacchi, J. M., and Pinto, W., Jr. (2005). Methylenetetrahydrofolate reductase gene
polymorphisms and their association with trisomy 21. Prenat. Diagn. 25, 11961199.
Hustad, S., Ueland, P. M., Vollset, S. E., Zhang, Y., Bjorke-Monsen, A. L., and Schneede, J.
(2000). Riboflavin as a determinant of plasma total homocysteine: Effect modification by
388 Philip Thomas and Michael Fenech
the methylenetetrahydrofolate reductase C677T polymorphism. Clin. Chem. 46,
10651071.
James, S. J., Pogribna, M., Pogribny, I. P., Melnyk, S., Hine, R. J., Gibson, J. B., Yi, P.,
Tafoya, D. L., Swenson, D. H., Wilson, V. L., and Gaylor, D. W. (1999). Abnormal
folate metabolism and mutation in the methylenetetrahydrofolate reductase gene may be
maternal risk factors for down syndrome. Am. J. Clin. Nutr. 70, 495501.
Kamiya, H., Kawakami, K., Miyanaga, T., Omura, K., Oda, M., Murakami, S., and
Watanabe, Y. (1998). A methylenetetrahydrofolate reductase polymorphism is associated
with expression of p16 in human lung cancer. Oncol. Rep. 5, 911914.
Kang, S. S., Wong, P. W., Susmano, A., Sora, J., Norusis, M., and Ruggie, N. (1991).
Thermolabile methylenetetrahydrofolate reductase: An inherited risk factor for coronary
artery disease. Am. J. Hum. Genet. 48, 536545.
Kawas, C. H. (2003). Clinical practice. Early Alzheimers disease. N. Engl. J. Med. 349,
10561063.
Keku, T., Millikan, R., Worley, K., Winkel, S., Eaton, A., Biscocho, L., Martin, C., and
Sandler, R. (2002). 5,10-Methylenetetrahydrofolate reductase codon 677 and 1298
polymorphisms and colon cancer in African Americans and whites. Cancer Epidemiol.
Biomarkers Prev. 11, 16111621.
Kim, J., Park, M. H., Kim, E., Han, C., Jo, S. A., and Jo, I. (2007). Plasma homocysteine is
associated with the risk of mild cognitive impairment in an elderly Korean population.
J. Nutr. 137, 20932097.
Kimura, M., Umegaki, K., Higuchi, M., Thomas, P., and Fenech, M. (2004). Methylene-
tetrahydrofolate reductase C677T polymorphism, folic acid and riboflavin are important
determinants of genome stability in cultured human lymphocytes. J. Nutr. 134, 4856.
Koch, M. C., Stegmann, K., Ziegler, A., Schroter, B., and Ermert, A. (1998). Evaluation of
the MTHFR C677T allele and the MTHFR gene locus in a German spina bifida
population. Eur. J. Pediatr. 157, 487492.
Kruman, I. I., Kumaravel, T. S., Lohani, A., Pedersen, W. A., Cutler, R. G., Kruman, Y.,
Haughey, N., Lee, J., Evans, M., and Mattson, M. P. (2002). Folic acid deficiency and
homocysteine impair DNA repair in hippocampal neurons and sensitize them to amyloid
toxicity in experimental models of Alzheimers disease. J. Neurosci. 22, 17521762.
Kupferminc, M. J., Eldor, A., Steinman, N., Many, A., Bar-Am, A., Jaffa, A., Fait, G., and
Lessing, J. B. (1999). Increased frequency of genetic thrombophilia in women with
complications of pregnancy. N. Engl. J. Med. 340, 913.
Laivuori, H., Kaaja, R., Ylikorkala, O., Hiltunen, T., and Kontula, K. (2000). 677 C !T
polymorphism of the methylenetetrahydrofolate reductase gene and preeclampsia. Obstet.
Gynecol. 96, 277280.
Le Marchand, L., Dolon, T., Hankin, J. H., Kolonel, L. N., Wilkens, L. R., and Seifried, A.
(2002). B-vitamin intake, metabolic genes, and colorectal cancer risk (United States).
Cancer Causes Control 13, 239248.
Leeda, M., Riyazi, N., de Vries, J. I., Jakobs, C., van Geijn, H. P., and Dekker, G. A. (1998).
Effects of folic acid and vitamin B6 supplementation on women with hyperhomocystei-
nemia and a history of preeclampsia or fetal growth restriction. Am. J. Obstet. Gynecol.
179, 135139.
Leopardi, P., Marcon, F., Caiola, S., Cafolla, A., Siniscalchi, E., Zijno, A., and Crebelli, R.
(2006). Effects of folic acid deficiency and MTHFR C677T polymorphism on sponta-
neous and radiation-induced micronuclei in human lymphocytes. Mutagenesis 21,
327333.
Lewis, S. J., Ebrahim, S., and Davey Smith, G. (2005). Meta-analysis of MTHFR 677C!T
polymorphism and coronary heart disease: Does totality of evidence support causal role
for homocysteine and preventive potential of folate? BMJ 331, 1053.
MTHFR, Common Polymorphisms, and Relation to Disease 389
Lievers, K. J., Boers, G. H., Verhoef, P., den Heijer, M., Kluijtmans, L. A., van der
Put, N. M., Trijbels, F. J., and Blom, H. J. (2001). A second common variant in the
methylenetetrahydrofolate reductase (MTHFR) gene and its relationship to MTHFR
enzyme activity, homocysteine, and cardiovascular disease risk. J. Mol. Med. 79,
522528.
Lucock, M., Yates, Z., Glanville, T., Leeming, R., Simpson, N., and Daskalakis, I. (2003).
A critical role for B-vitamin nutrition in human developmental and evolutionary
biology. Nutr. Res. 23, 14631475.
Ma, J., Stampfer, M. J., Giovannucci, E., Artigas, C., Hunter, D. J., Fuchs, C.,
Willett, W. C., Selhub, J., Hennekens, C. H., and Rozen, R. (1997). Methylenetetra-
hydrofolate reductase polymorphism, dietary interactions, and risk of colorectal cancer.
Cancer Res. 57, 10981102.
Mattson, M. P. (2004). Pathways towards and away from Alzheimers disease. Nature 430,
631639.
Mattson, M. P., Keller, J. N., and Begley, J. G. (1998). Evidence for synaptic apoptosis. Exp.
Neurol. 153, 3548.
McCaddon, A., Davies, G., Hudson, P., Tandy, S., and Cattell, H. (1998). Total serum
homocysteine in senile dementia of Alzheimer type. Int. J. Geriatr. Psychiatry 13,
235239.
Morita, H., Kurihara, H., Tsubaki, S., Sugiyama, T., Hamada, C., Kurihara, Y., Shindo, T.,
Oh-hashi, Y., Kitamura, K., and Yazaki, Y. (1998). Methylenetetrahydrofolate reductase
gene polymorphism and ischemic stroke in Japanese. Arterioscler Thromb. Vasc. Biol. 18,
14651469.
Morris, M. S. (2003). Homocysteine and Alzheimers disease. Lancet Neurol. 2, 425428.
Morrison, K., Papapetrou, C., Hol, F. A., Mariman, E. C., Lynch, S. A., Burn, J., and
Edwards, Y. H. (1998). Susceptibility to spina bifida: An association study of five
candidate genes. Ann. Hum. Genet. 62, 379396.
Munoz-Moran, E., Dieguez-Lucena, J. L., Fernandez-Arcas, N., Peran-Mesa, S., and
Reyes-Engel, A. (1998). Genetic selection and folate intake during pregnancy. Lancet
352, 11201121.
Ramsbottom, D., Scott, J. M., Molloy, A., Weir, D. G., Kirke, P. N., Mills, J. L.,
Gallagher, P. M., and Whitehead, A. S. (1997). Are common mutations of cystathionine
beta-synthase involved in the aetiology of neural tube defects? Clin. Genet. 51, 3942.
Rasnick, D., and Duesberg, P. H. (1999). How aneuploidy affects metabolic control and
causes cancer. Biochem. J. 340(Pt 3), 621630.
Ray, J. G., and Laskin, C. A. (1999). Folic acid and homocyst(e)ine metabolic defects and the
risk of placental abruption, pre-eclampsia and spontaneous pregnancy loss: A systematic
review. Placenta 20, 519529.
Refsum, H., and Ueland, P. M. (1998). Recent data are not in conflict with homocysteine as
a cardiovascular risk factor. Curr. Opin. Lipidol. 9, 533539.
Rivlin, R. S. (1996). In Present Knowledge in Nutrition 7th. ed., pp. 206219.
Rozen, R. (1997). Genetic predisposition to hyperhomocysteinemia: Deficiency of methy-
lenetetrahydrofolate reductase (MTHFR). Thromb. Haemost. 78, 523526.
Scarpa, S., Fuso, A., DAnselmi, F., and Cavallaro, R. A. (2003). Presenilin 1 gene silencing
by S-adenosylmethionine: A treatment for Alzheimer disease? FEBS Lett. 541, 145148.
Schneider, J. A., Rees, D. C., Liu, Y. T., and Clegg, J. B. (1998). Worldwide distribution of
a common methylenetetrahydrofolate reductase mutation. Am. J. Hum. Genet. 62,
12581260.
Sellers, T. A., Kushi, L. H., Cerhan, J. R., Vierkant, R. A., Gapstur, S. M., Vachon, C. M.,
Olson, J. E., Therneau, T. M., and Folsom, A. R. (2001). Dietary folate intake, alcohol,
and risk of breast cancer in a prospective study of postmenopausal women. Epidemiology
12, 420428.
390 Philip Thomas and Michael Fenech
Seshadri, S., Beiser, A., Selhub, J., Jacques, P. F., Rosenberg, I. H., DAgostino, R. B.,
Wilson, P. W., and Wolf, P. A. (2002). Plasma homocysteine as a risk factor for dementia
and Alzheimers disease. N. Engl. J. Med. 346, 476483.
Shaw, G. M., Rozen, R., Finnell, R. H., Wasserman, C. R., and Lammer, E. J. (1998).
Maternal vitamin use, genetic variation of infant methylenetetrahydrofolate reducatase,
and risk for spina bifida. Am. J. Epidemiol. 148, 3037.
Shea, T. B., and Rogers, E. (2002). Homocysteine and dementia. N. Engl. J. Med. 346,
2007; author reply 2008..
Sibani, S., Christensen, B., OFerrall, E., Saadi, I., Hiou-Tim, F., Rosenblatt, D. S., and
Rozen, R. (2000). Characterization of six novel mutations in the methylenetetrahydro-
folate reductase (MTHFR) gene in patients with homocystinuria. Hum. Mutat. 15,
280287.
Sibani, S., Leclerc, D., Weisberg, I. S., OFerrall, E., Watkins, D., Artigas, C.,
Rosenblatt, D. S., and Rozen, R. (2003). Characterization of mutations in severe
methylenetetrahydrofolate reductase deficiency reveals an FAD-responsive mutation.
Hum. Mutat. 21, 509520.
Skibola, C. F., Smith, M. T., Kane, E., Roman, E., Rollinson, S., Cartwright, R. A., and
Morgan, G. (1999). Polymorphisms in the methylenetetrahydrofolate reductase gene are
associated with susceptibility to acute leukemia in adults. Proc. Natl. Acad. Sci. 96,
1281012815.
St George-Hyslop, P. H. (2000). Piecing together Alzheimers. Sci. Am. 283, 7683.
Steegers-Theunissen, R. P., Boers, G. H., Trijbels, F. J., Finkelstein, J. D., Blom, H. J.,
Thomas, C. M., Borm, G. F., Wouters, M. G., and Eskes, T. K. (1994). Maternal
hyperhomocysteinemia: A risk factor for neural-tube defects? Metabolism 43, 14751480.
Steegers-Theunissen, R. P., Boers, G. H., Blom, H. J., Nijhuis, J. G., Thomas, C. M.,
Borm, G. F., and Eskes, T. K. (1995). Neural tube defects and elevated homocysteine
levels in amniotic fluid. Am. J. Obstet. Gynecol. 172, 14361441.
Stern, L. L., Mason, J. B., Selhub, J., and Choi, S. W. (2000). Genomic DNA hypomethyla-
tion, a characteristic of most cancers, is present in peripheral leukocytes of individuals
who are homozygous for the C677T polymorphism in the methylenetetrahydrofolate
reductase gene. Cancer Epidemiol. Biomarkers Prev. 9, 849853.
Suzuki, T., Fujii, M., and Ayusawa, D. (2002). Demethylation of classical satellite 2 and 3
DNA with chromosomal instability in senescent human fibroblasts. Exp. Gerontol. 37,
10051014.
Thomas Kaiser, S. P. B., and Moses, E. K. (2000). Methylenetetrahydrofolate reductase
polymorphisms are not a risk factor for pre-eclampsia/eclampsia in Australian women.
Gynecol. Obstet. Invest. 50, 100102.
Trembath, D., Sherbondy, A. L., Vandyke, D. C., Shaw, G. M., Todoroff, K.,
Lammer, E. J., Finnell, R. H., Marker, S., Lerner, G., and Murray, J. C. (1999). Analysis
of select folate pathway genes, PAX3, and human T in a midwestern neural tube defect
population. Teratology 59, 331341.
Ueland, P. M., Hustad, S., Schneede, J., Refsum, H., and Vollset, S. E. (2001). Biological
and clinical implications of the MTHFR C677T polymorphism. Trends Pharmacol. Sci.
22, 195201.
van der Put, N. M., Gabreels, F., Stevens, E. M., Smeitink, J. A., Trijbels, F. J., Eskes, T. K.,
van den Heuvel, L. P., and Blom, H. J. (1998). A second common mutation in the
methylenetetrahydrofolate reductase gene: An additional risk factor for neural-tube
defects? Am. J. Hum. Genet. 62, 10441051.
Viel, A., DallAgnese, L., Simone, F., Canzonieri, V., Capozzi, E., Visentin, M. C.,
Valle, R., and Boiocchi, M. (1997). Loss of heterozygosity at the 5,10-methylenetetra-
hydrofolate reductase locus in human ovarian carcinomas. Br. J. Cancer 75, 11051110.
MTHFR, Common Polymorphisms, and Relation to Disease 391
Wagner, C. (1995). Biochemical role of folate in cellular metabolism. In Folate in Health
and Disease (L. B. Bailey, ed.), p. 26, CRC press, Boca Raton, Florida, USA.
Wakutani, Y., Kowa, H., Kusumi, M., Nakaso, K., Yasui, K., Isoe-Wada, K., Yano, H.,
Urakami, K., Takeshima, T., and Nakashima, K. (2004). A haplotype of the methyle-
netetrahydrofolate reductase gene is protective against late-onset Alzheimers disease.
Neurobiol. Aging 25, 291294.
Wald, D. S., Law, M., and Morris, J. K. (2002). Homocysteine and cardiovascular disease:
Evidence on causality from a meta-analysis. BMJ 325, 12021206.
Wang, H. X., Wahlin, A., Basun, H., Fastbom, J., Winblad, B., and Fratiglioni, L. (2001).
Vitamin B(12) and folate in relation to the development of Alzheimers disease. Neurology
56, 11881194.
Wang, X., Thomas, P., Xue, J., and Fenech, M. (2004). Folate deficiency induces aneu-
ploidy in human lymphocytes in vitro-evidence using cytokinesis-blocked cells and
probes specific for chromosomes 17 and 21. Mutat. Res. 551, 167180.
Weisberg, I., Tran, P., Christensen, B., Sibani, S., and Rozen, R. (1998a). A second genetic
polymorphism in methylenetetrahydrofolate reductase (MTHFR) associated with
decreased enzyme activity. Mol. Genet. Metab. 64, 169172.
Weisberg, I., Tran, P., Christensen, B., Sibani, S., and Rozen, R. (1998b). A second genetic
polymorphism in methylenetetrahydrofolate reductase (MTHFR) associated with
decreased enzyme activity. Mol. Genet. Metab. 64, 169172.
Wilcken, B., Bamforth, F., Li, Z., Zhu, H., Ritvanen, A., Redlund, M., Stoll, C.,
Alembik, Y., Dott, B., Czeizel, A. E., Gelman-Kohan, Z., Scarano, G., et al. (2003).
Geographical and ethnic variation of the 677C > T allele of 5,10 methylenetetrahydro-
folate reductase (MTHFR): Findings from over 7000 newborns from 16 areas world
wide. J. Med. Genet. 40, 619625.
Williams, J. H., Pereira, E. A., Budge, M. M., and Bradley, K. M. (2002). Minimal
hippocampal width relates to plasma homocysteine in community-dwelling older people.
Age Ageing 31, 440444.
Zetterberg, H., Zafiropoulos, A., Spandidos, D. A., Rymo, L., and Blennow, K. (2003).
Genegene interaction between fetal MTHFR 677C>T and transcobalamin 776C>G
polymorphisms in human spontaneous abortion. Hum. Reprod. 18, 19481950.
Zhang, S., Hunter, D. J., Hankinson, S. E., Giovannucci, E. L., Rosner, B. A.,
Colditz, G. A., Speizer, F. E., and Willett, W. C. (1999). A prospective study of folate
intake and the risk of breast cancer. JAMA 281, 16321637.
Zingg, J. M., and Jones, P. A. (1997). Genetic and epigenetic aspects of DNA methylation
on genome expression, evolution, mutation and carcinogenesis. Carcinogenesis 18,
869882.
392 Philip Thomas and Michael Fenech
C H A P T E R F O U R T E E N
Mitochondrial
Methylenetetrahydrofolate
Dehydrogenase,
Methenyltetrahydrofolate
Cyclohydrolase, and
Formyltetrahydrofolate Synthetases
Karen E. Christensen* and Robert E. MacKenzie
Contents
I. Introduction 394
II. Yeast Mitochondria Contain a Trifunctional
DehydrogenaseCyclohydrolaseSynthetase 395
III. Mammalian Mitochondrial
Methylenetetrahydrofolate Dehydrogenase 395
A. Discovery and characterization of NAD-dependent
methylenetetrahydrofolate dehydrogenasecyclohydrolase
(MTHFD2) 395
B. Differences between the NAD- and NADP-dependent
dehydrogenasecyclohydrolases 396
C. Distribution and expression of methylenetetrahydrofolate
dehydrogenasecyclohydrolases 401
IV. Mitochondrial Formyltetrahydrofolate Synthetase 403
A. Discovery 403
B. Characterization of the monofunctional synthetase 404
C. Expression and distribution 404
V. Conclusion 405
A. Metabolic compartmentation of the folate enzymes 405
B. Functions of mitochondria in folate-mediated metabolism 406
C. Mathematical models 407
References 408
Vitamins and Hormones, Volume 79
#
2008 Elsevier Inc.
ISSN 0083-6729, DOI: 10.1016/S0083-6729(08)00414-7 All rights reserved.
* Montreal Childrens Hospital Research Institute, Montreal, QC, Canada H3Z 2Z3
{
Department of Biochemistry, McGill University, Montreal, QC, Canada H3G 1Y6
393
Abstract
Folate-mediated metabolism involves enzyme-catalyzed reactions that occur in
the cytoplasmic, mitochondrial, and nuclear compartments in mammalian cells.
Which of the folate-dependent enzymes are expressed in these compartments
depends on the stage of development, cell type, cell cycle, and whether or not
the cell is transformed. Mitochondria become formate-generating organelles in
cells and tissues expressing the MTHFD2 and MTHFD1L genes. The products of
these nuclear genes were derived from trifunctional precursor proteins, expres-
sing methylenetetrahydrofolate dehydrogenasemethenyltetrahydrofolate
cyclohydrolase, and formyltetrahydrofolate synthetase activities. The MTHFD2
protein is a bifunctional protein with dehydrogenase and cyclohydrolase activ-
ities that arose from a trifunctional precursor through the loss of the synthetase
domain and a novel adaptation to NAD rather than NADP specificity for the
dehydrogenase. The MTHFD1L protein retains the size of its trifunctional precur-
sor, but through the mutation of critical residues, both the dehydrogenase and
cyclohydrolase activities have been silenced. MTHFD1L is thus a monofunctional
formyltetrahydrofolate synthetase. This review discusses the properties and
functions of these mitochondrial proteins and their role in supporting cytosolic
purine synthesis during embryonic development and in cells undergoing rapid
growth. 2008 Elsevier Inc.
I. Introduction
Folate-mediated metabolism is a complex series of pathways that
involve several amino acids, the synthesis of purines and thymidylate, the
support of cellular methylation reactions via recycling of homocysteine to
methionine, and the synthesis of S-adenosylmethionine. Serine, glycine,
and formate are the main sources of the one-carbon folates used in these
biosynthetic pathways. Folate-dependent enzymes are found in different
subcellular compartments, often with the same activity appearing in more
than one compartment. In this regard, there are significant differences seen
between species (Christensen and MacKenzie, 2006). The object of this
review is to concentrate on the enzyme activities found in mammalian
mitochondria, and to discuss possible advantages of this compartmentation.
The cytoplasm and mitochondria of mammalian cells contain isoforms of
serine hydroxymethyl transferase (SHMT1 and SHMT2, respectively) that
generate glycine and methylenetetrahydrofolate from serine, in a reversible
reaction. In essence, it is the fate of this methylenetetrahydrofolate that is the
key to understanding the contribution of the mitochondria to overall folate-
mediated metabolism. The genetics and biochemical characterization of
these enzymes in yeast mitochondria provide an informative introduction
to the situation in mammals.
394 Karen E. Christensen and Robert E. MacKenzie
II. Yeast Mitochondria Contain a Trifunctional
DehydrogenaseCyclohydrolaseSynthetase
Early genetic evidence indicated that three enzyme activities, NADP-
dependent methylenetetrahydrofolate dehydrogenase, methenyltetrahydrofo-
late cyclohydrolase, and 10-formyltetrahydrofolate synthetase were physically
associated in yeast ( Jones, 1977). Purification of these enzymes first from liver
(Paukert et al., 1976; Tan et al., 1977) and then fromyeast (Paukert et al., 1977)
demonstrated that they actually form a trifunctional protein in eukaryotes.
Using proteolysis, several laboratories were able to show that the eukaryotic
trifunctional enzymes were composed of an amino terminal dehydrogenase
cyclohydrolase domain attached to a synthetase domain (MacKenzie, 1984).
Expression of the human trifunctional enzyme (MTHFD1) and its two active
domains in Escherichia coli (Humand MacKenzie, 1991) allowed the character-
ization of the linker region connecting the domains. The cytoplasmic form
of the yeast dehydrogenasecyclohydrolasesynthetase (ADE3) was cloned by
Staben and Rabinowitz (1986). The mitochondrial isozyme (MIS1) was
subsequently cloned and characterized by Shannon and Rabinowitz (1986
and 1988). This distribution of two NADP-dependent trifunctional
dehydrogenasecyclohydrolasesynthetases in yeast led to the proposal of a
similar model for mammalian cells (Appling, 1991). In the yeast model, one-
carbonunits derivedfromserine inmitochondria as methylenetetrahydrofolate
are oxidized to formyltetrahydrofolate and, by the reverse activity of the
formyltetrahydrofolate synthetase, produce formate. The formate released
from the mitochondria can then be incorporated into cytoplasmic forms of
one-carbon tetrahydrofolates by the activities of the cytoplasmic ADE3. This
allows the mitochondria to contribute one-carbon units for use in biosynthetic
reactions in the cytoplasm.
III. Mammalian Mitochondrial
Methylenetetrahydrofolate Dehydrogenase
A. Discovery and characterization of NAD-dependent
methylenetetrahydrofolate
dehydrogenasecyclohydrolase (MTHFD2)
An observation by Scrimgeour and Huennekens (1960) that extracts of
Ehrlich ascites tumor cells contained NAD-dependent methylenetetrahydro-
folate dehydrogenase activity went unsubstantiateduntil Mejia andMacKenzie
(1985) demonstrated that the same activity is detectable in all transformed and
nondifferentiated cells assayed. Moreover, the NAD-dependent activity could
DehydrogenaseCyclohydrolaseSynthetase 395
be separated from the known cytosolic NADP-dependent dehydrogenase by
column chromatography. At the time these authors suggested that the use of
NAD rather than NADP could shift the cellular one-carbon folate pool in
favor of elevated formyltetrahydrofolate levels. This shift would favor the
synthesis of purines rather than methionine. However, the metabolic role of
MTHFD2 has been clarified only recently as discussed in Section V.B.
Upon purification, the protein was found to be a bifunctional NAD-
dependent methylenetetrahydrofolate dehydrogenasecyclohydrolase with
a polypeptide size of 34,000 Da, rather than the 100,000 Da of the known
NADP-dependent dehydrogenasecyclohydrolasesynthetase enzymes (Mejia
et al., 1986). In addition, it required Mg
2
(or Mn
2
) for activity, unlike
any known dehydrogenase. Subcellular fractionation of ascites tumor cells
demonstrated that MTHFD2 is located in mitochondria (Mejia and
MacKenzie, 1988). This was confirmed by the presence of a mitochondrial
targeting sequence in the amino acid sequence deduced from the cDNA clone
(Belanger and MacKenzie, 1989). It was first shown that the divalent ion is
required for NADbinding (Rios-Orlandi and MacKenzie, 1988), and later and
more surprisingly, that both Mg
2
and inorganic phosphate are absolutely
essential for this purpose (Yang and MacKenzie, 1993). These authors found
that the dehydrogenase can also use NADP but with a higher K
m
and lower
V
max
. In this case, inorganic phosphate is not only not required, but also is a
competitive inhibitor against NADP. These properties led to the proposal that
the bifunctional enzyme evolved from an NADP-dependent precursor and is
the mammalian homolog of the mitochondrial yeast MIS1 protein.
B. Differences between the NAD- and NADP-dependent
dehydrogenasecyclohydrolases
Characterization of the dehydrogenasecyclohydrolase activities of the
cytosolic enzyme showed that they are not catalytically independent
because NADP is an inhibitor of the cyclohydrolase (Smith and
MacKenzie, 1983), and most convincingly, that about half of the methe-
nyltetrahydrofolate produced by the dehydrogenase (5060%) is prefer-
entially channeled through the cyclohydrolase rather than being released
into the bulk medium (Cohen and MacKenzie, 1978). Pawelek and
MacKenzie (1998) expressed the cDNA clones for three enzymes with
different cofactor specificities: the DC domain of the human MTHFD1
(DC301), the bifunctional DC enzyme from Photobacterium phosphoreum,
and the human MTHFD2. The rates of forward and reverse dehydrogenase
and cyclohydrolase activities as well as the overall forward and reverse
activities were measured. The kinetic analysis revealed that the overall
forward reaction from methylene- to formyltetrahydrofolate shows a
hydride transfer kinetic isotope effect, whereas the overall reverse reaction
396 Karen E. Christensen and Robert E. MacKenzie
does not. This clearly showed that the dehydrogenase and cyclohydrolase
catalyzed reactions cannot be considered as a single complex reaction. It also
demonstrated that the rate of the overall reverse reaction of formyl- to
methylene conversion is the same as that for the first step, catalyzed by the
cyclohydrolase. This showed that the channeling of the intermediate methe-
nyltetrahydrofolate in the reverse direction is 100% efficient. Moreover, it
was found that the presence of 2
0
,5
0
-ADP, an analog of NADP, stimulated the
reverse cyclohydrolase activity of the human MTHFD1 DCdomain by more
than twofold. The cyclohydrolase activity of MTHFD2 could not be stimu-
lated by 2
0
,5
0
-ADP or by 5
0
-ADP. This suggests that the NADP-dependent
DC domains are optimized to catalyze the reverse direction by efficiently
retaining the labile methenyl intermediate and converting it to methylenete-
trahydrofolate. In contrast, the properties of the NAD-dependent DCsuggest
that it is not optimized for catalysis of the reverse reaction.
The X-ray structure of the NADP-dependent DC domain of MTHFD1
(Allaire et al., 1998) as well as structures with bound inhibitors (Schmidt
et al., 2000 and Fig. 14.1) set the stage for site-directed mutagenesis
Figure 14.1 The methylenetetrahydrofolate dehydrogenasemethenyltetrahydrofolate
cyclohydrolase binding site of MTHFD1 (Protein Data Bank ID: 1DIB; Schmidt et al.,
2000). The folate analog LY345899 is shown in green; NADP is shown in yellow.
Arg173 and Ser197 are involved in NADP binding and Asp125 is required for binding
the folate substrate. Lys56 and Gln100 are involved in catalyzing the cyclohydrolase
reaction. Figure generated using PyMol version 0.96 (DeLano, 2004).
DehydrogenaseCyclohydrolaseSynthetase 397
experiments to understand the binding of the NADP (Pawelek et al., 2000)
and the kinetic mechanism of the dehydrogenasecyclohydrolase activities
catalyzed by this bifunctional site (Sundararajan and MacKenzie, 2002).
These studies showed that Lys56, supported by its interaction with
Gln100, provides the acidbase catalysis required for the cyclohydrolase.
A double mutant K56Q/Q100K has no cyclohydrolase activity but retains
two-thirds of the normal dehydrogenase activity. Asp125 is required for the
binding of the folate substrates; both activities are abolished by mutations at
this position. The binding of NADP involves Arg173, with a minor
contribution from Ser197. Mutation of Arg173 causes a 500-fold increase
in the K
m
for NADP, while mutation of Ser197 causes a 20-fold increase.
The results showed very clearly that the majority of the interaction of the
NADP with the enzyme is via the 2
0
-phosphate (Fig. 14.1).
The crystal structure and the characterization of NADP binding to the
cytoplasmic enzyme were critical to allow understanding of the mitochon-
drial MTHFD2. In the absence of X-ray structures, Christensen et al.
(2005a) built a homology model of MTHFD2 based on three related crystal
structures. Yang and MacKenzie (1993) found that P
i
is a competitive
inhibitor when NADP is used as a substrate for the NAD-dependent
methylenetetrahydrofolate dehydrogenase. This was interpreted to mean
that P
i
binds to MTHFD2 in a position similar to that of the 2
0
-phosphate of
NADP in the DC domain of MTHFD1. The homology model predicted
that Arg166 and Arg198 are located in the region of the 2
0
-hydroxyl of
bound NAD. Characterization of mutants showed that Arg166 is critically
responsible for P
i
binding, and is positioned properly by its interaction with
Asp190. Arg166 plays a role in phosphate binding similar to that of Arg173
in MTHFD1 and cannot be mutated without complete loss of the dehy-
drogenase activity. Because mutants of Arg198 retain some dehydrogenase
activity, it was proposed that it assists in P
i
binding and proper orientation
and plays a role similar to that of Ser197 in MTHFD1. Thus in MTHFD2,
the P
i
is retained in a positively charged pocket where it interacts with the
2
0
-hydroxyl of bound NAD (Fig. 14.2).
The mitochondrial enzyme, MTHFD2, also absolutely requires Mg
2
for activity. Possible binding sites were sought by examination of multiple
sequence alignments for Asp and Glu residues conserved in NAD-depen-
dent DC enzymes but not in NADP-dependent enzymes. The identified
residues were tested by mutagenesis and only Asp133 was found to be
essential for activity; D133E retains no dehydrogenase activity showing
that positioning of the carboxyl group is critical. Mg
2
binds to a cavity
bounded by Asp133, inorganic phosphate, and the NAD cofactor, and
assists in the binding and positioning of the P
i
, similar to the role assigned
to Arg198. Thus, this enzyme uses Mg
2
and P
i
to bind NAD in a manner
normally found with NADP (Fig. 14.2).
398 Karen E. Christensen and Robert E. MacKenzie
Comparison of the structures of NAD- and NADP-dependent dehy-
drogenases (Fig. 14.3) pointed out that the binding of NADP to the DC
domain of MTHFD1 was almost entirely explained by interaction of
arginine residues with the 2
0
-phosphate of the molecule. In comparison,
NAD-binding proteins, such as alcohol dehydrogenase, exhibit multiple
interactions between the cofactor and protein with considerably more com-
plementarity of fit. The NAD-binding site of MTHFD2 shows the overall
looser interaction with the cofactor that is characteristic of NADP sites. To
convert this site to an NAD-binding site would have entailed the develop-
ment of enhanced complementarity and the development of multiple
cofactorprotein interactions. Instead, through evolution, Mg
2
and P
i
have been used in a novel fashion to mimic the binding of NADP, effectively
converting this enzyme to NAD specificity.
The mode of binding of NAD, indicating that the precursor of this protein
was NADP specific, fits nicely with earlier work. It was recognized from the
isolation of the cDNA encoding MTHFD2 (Belanger and MacKenzie, 1989)
that the message for the protein had a rather long 3
0
-untranslated region. Patel
et al. (2002) isolated the cDNA and the gene encoding the murine MTHFD1
and compared these with the cDNA of MTHFD2. They found DNA
sequence homologies between the synthetase region of MTHFD1 and the
Figure 14.2 The ion-binding site of MTHFD2. Arg166 (assisted by Asp190) and
Arg198 bind the inorganic phosphate. Asp133, inorganic phosphate, and NAD form
the binding site for the magnesium ion. This binding site positions the phosphate and
magnesium ions so that they form ionic and hydrogen bonds with NAD, facilitating
cofactor binding. Figure generated using PyMol version 0.96 (DeLano, 2004). Adapted
from Christensen et al. (2005a, Fig. 4, p. 34321).
DehydrogenaseCyclohydrolaseSynthetase 399
Figure 14.3 Comparison of the cofactor-binding sites of (A) alcohol dehydrogenase (Protein Data Bank ID: 1HDX), (B) MTHFD2, and
(C) the DC domain of the human MTHFD1 (Protein Data Bank ID: 1DIB). There is a close complementarity of fit between the NAD
cofactor and the protein in classic NAD-dependent proteins, such as alcohol dehydrogenase, that permits multiple cofactorprotein
interactions to bind the cofactor. This close fit is disrupted in MTHFD2; the NAD-binding site is more similar to a classic NADP-binding
site, such as that of MTHFD1. NADP-binding sites usually depend on interactions with the 2
0
-phosphate of NADP for cofactor binding. In
MTHFD2, magnesium and inorganic phosphate ions in the NAD-binding site set up a web of ionic and hydrogen bonding interactions that
compensate for the lack of a covalent bond with phosphate. This allows the adaptation of an NADP-binding site to bind NAD.
Figure generated using PyMol version 0.96 (DeLano, 2004). From Christensen et al. (2005a, Fig. 5, p. 34322).
3
0
-untranslated region of MTHFD2. Similar observations were made using
human and Drosophila sequences. It is abundantly clear that the mitochondrial
NAD-dependent enzyme did not arise from a bifunctional prokaryotic-
like bifunctional enzyme, but from an NADP-dependent methylenetetrahy-
drofolate dehydrogenasecyclohydrolasesynthetase through loss of the
synthetase domain.
C. Distribution and expression of methylenetetrahydrofolate
dehydrogenasecyclohydrolases
Folate metabolism is compartmentalized in mammalian cells (Fig. 14.4),
primarily between the cytoplasm and mitochondria (Christensen and
MacKenzie, 2006). However, recent exciting work by Woeller et al.
(2007) demonstrates that the three folate-dependent activities involved in
de novo thymidylate synthesis contain small ubiquitin-like modifier (SUMO)
modification consensus sequences and are located in the nucleus during S
and G2/M phases of the cell cycle in MCF-7 cells. Clearly, the expression
and localization of the folate-dependent enzymes is dynamic, dependent on
tissue, cell type, and cell cycle.
5-MethylTHF
Methionine
Thymidylate
Purines
Mitochondria
MethenylTHF
Serine
MethyleneTHF
FormylTHF
Formate
THF
Glycine
D
NAD
NADH
Mg
+2
+ Pi
C
MethenylTHF
Serine
MethyleneTHF
FormylTHF
Formate
THF
Glycine
NADP
NADPH
D
C
S
ADP + Pi
ATP THF
S
ADP + Pi
ATP
THF
Cytoplasm
Figure 14.4 The cytoplasmic and mitochondrial folate pathways of embryonic and
transformed cells. In the cytoplasm, the methylenetetrahydrofolate dehydrogenase (D),
methenyltetrahydrofolate cyclohydrolase (C), and formyltetrahydrofolate synthetase
(S) activities are found in the trifunctional protein MTHFD1. In the mitochondria of
embryonic and transformed cells, the dehydrogenase and cyclohydrolase activities are
found in MTHFD2 whereas the synthetase activity is found in MTHFD1L. From
Christensen et al. (2005b, Fig. 8, p. 7601).
DehydrogenaseCyclohydrolaseSynthetase 401
The cytoplasmic MTHFD1 is ubiquitously expressed in all mammalian
cells although there is extensive variation in the level of mRNA found in
different tissues of the mouse (Peri and MacKenzie, 1991) and in the rat
(Thigpen et al., 1990). Despite its tissue-specific variation, it is probably a
house-keeping enzyme. The overall reaction (including cyclohydrolase):
methylenetetrahydrofolate NADP
,formyltetrahydrofolate NADPH
has a K
eq
16 and the dehydrogenasecyclohydrolase appears to act
to maintain these cytoplasmic pools in equilibrium (Pelletier and
MacKenzie, 1995).
In contrast, the mitochondrial MTHFD2 has been detected only in
transformed cell lines and in embryonic tissues; it has not been detected in
differentiated tissues (Mejia and MacKenzie, 1985; Smith et al., 1990).
Although no NAD-dependent dehydrogenase activity is detectable in
adult tissue extracts, very small amounts of mRNA can be detected (Peri
and MacKenzie, 1993). The MTHFD2 gene is induced in quiescent Balb/c
3T3 cells by serum or phorbol esters whereas expression of MTHFD1 is
unaffected by mitogenic signals (Peri and MacKenzie, 1991).
The limited expression distribution pattern observed for the mitochondrial
MTHFD2 raised questions as to its role and importance. Di Pietro et al. (2002)
performed a knockout of the nuclear gene encoding this protein in mice and
found that the null mutation was embryonic lethal at about 12 days gestation.
In particular, it was striking that the normal development of hematogenesis in
the liver did not occur. While establishing a critical role for this protein, the
results suggested that perhaps the expression is liver-specific during embryo-
genesis. That this is not the case was established by in situ hybridization studies
that showed generalized transcription of the mRNA in all tissues of embryos
during early development (Di Pietro et al., 2004).
Characterization of null fibroblast cell lines established from MTHFD2
knockout embryos established that they contain functional mitochondria, but
unlike wild-type cells, are glycine auxotrophs (Patel et al., 2003). This
observation shows that the cytoplasmic SHMT1 alone cannot provide the
cellular requirement for glycine. Without MTHFD2, the accumulation of
mitochondrially generated methylenetetrahydrofolate blocks production of
glycine by SHMT2 (Patel et al., 2003). These authors also demonstrated that
the null cells preferentially incorporate exogenously added formate into
purines when compared to controls, indicating that less endogenous formate
is being produced by the mitochondria in these cells. These experiments
indicate a role for mitochondria in providing both glycine and formate to the
cell during times of rapid growth. Although the presence of MTHFD2
explained the production of mitochondrial formyltetrahydrofolate, it was
402 Karen E. Christensen and Robert E. MacKenzie
not clear from these studies how the formyltetrahydrofolate yielded formate
for use in the cytoplasm.
IV. Mitochondrial Formyltetrahydrofolate
Synthetase
A. Discovery
The distribution of both cytoplasmic and mitochondrial trifunctional
enzymes in yeast provided a likely model for mammalian cells. Barlowe
and Appling (1988), in a detailed investigation, isolated mitochondria from
rat liver and detected methylenetetrahydrofolate dehydrogenase and for-
myltetrahydrofolate synthetase activities in the most highly purified frac-
tions. Their work established the concept in mammalian metabolism that
mitochondria can supply one-carbon units for cytoplasmic use. While
mitochondrial preparations have these activities, the amounts are low, and
it was not clear that the story was complete in the absence of either a
purified protein or identification of a gene encoding a mitochondrial
isoform of MTHFD1. Moreover, the presence of MTHFD2 would put
both NAD- and NADP-dependent dehydrogenases in the same metabolic
compartment. With the high ratio of NAD/NADH and low ratio of
NADP/NADPH, these enzymes would be catalyzing a futile cycle if both
were present in the same metabolic compartment. However, the expression
of MTHFD2 (see Section III.C) indicated that it is not found in adult
tissues, leaving open the possibility that two MTHFD1 isoforms might
exist in mitochondria, but expressed at different times.
One possible argument against the yeast model holding true for mam-
mals was that mitochondrial preparations might contain small amounts of
MTHFD1, and that in fact, there was no mitochondrial trifunctional
isoform in mammalian cells. Christensen et al. (2005b) addressed this ques-
tion by undertaking to inactivate the MTHFD1 gene in embryonic stem
cells. The null embryonic stem cells were used to generate spontaneously
immortalized fibroblast cell lines. These cell lines were found to be purine
auxotrophs, as would be predicted from the loss of both the dehydrogenase
and synthetase activities that provide formyltetrahydrofolate for purine
synthesis. In the absence of the cytoplasmic dehydrogenase, cyclohydrolase,
and synthetase activities, identical enzyme activities of the products of
other genes could be detected unambiguously. Extracts of these cells had
no NADP-dependent methylenetetrahydrofolate dehydrogenase, but did
exhibit detectable 10-formyltetrahydrofolate synthetase activity that was
localized to the mitochondria.
While these results of Christensen et al. (2005b) supported the presence
of a monofunctional synthetase in mitochondria, no such protein had been
DehydrogenaseCyclohydrolaseSynthetase 403
isolated. However, Prasannan et al. (2003) made the important discovery of
the existence of a human gene encoding a mitochondrial isoform of
MTHFD1 (MTHFD1L). The protein encoded contains 978 amino acids,
including a 60 amino acid N-terminal extension containing a mitochondrial
targeting sequence, with a predicted cleavage site between residues 31 and 32.
Transfection experiments with the complete cDNA in CHO cells confirmed
that the protein is targeted to mitochondria. The length of the protein is
consistent with it being a trifunctional DCS; expression of a cDNA clone
in yeast demonstrated formyltetrahydrofolate synthetase activity, but not
methylenetetrahydrofolate dehydrogenase activity.
How could the existence of MTHFD1L be reconciled with the presence
of a putative monofunctional synthetase uncovered by Christensen et al.
(2005b)? Considerable evidence is available on residues necessary for the
methylenetetrahydrofolate dehydrogenase and cyclohydrolase activities
(Pawelek et al., 2000; Sundararajan and MacKenzie, 2002). By comparison
of multiple sequence alignments, it was found that MTHFD1L contains a
nonconservative mutation in a lysine residue critical for cyclohydrolase
activity, as well as in three residues critical for the binding of NADP,
thereby supporting a prediction that this enzyme would exhibit only the
synthetase activity (Christensen et al., 2005b). These authors suggested that
the inactive DC domain would still dimerize as seen in the structure of the
DC domain of MTHFD1 (Allaire et al., 1998), and thus serve to maintain
the dimeric interaction of the MTHFD1L enzyme.
B. Characterization of the monofunctional synthetase
Walkup and Appling (2005) expressed the cDNA encoding MTHFD1L in
E. coli, and characterized the purified protein. They confirmed that
MTHFD1L is monofunctional, exhibiting only synthetase activity. Gel
filtration studies and chemical crosslinking demonstrated that the protein
is a dimer. Kinetic analysis showed that the synthetase exhibits increased
efficiency with the polyglutamate forms of tetrahydrofolate, consistent with
the properties of other eukaryotic and prokaryotic enzymes; the K
cat
/K
m
THF values were 0.03 s
1
M
1
for the monoglutamate and 0.71 s
1
M
1
for the pentaglutamate substrates.
The sequential activities of MTHFD2 and MTHFD1L enable the pro-
duction of formate in mitochondria from methylenetetrahydrofolate gener-
ated either from serine via SHMT2 or from glycine via the glycine cleavage
system.
C. Expression and distribution
MTHFD1L is similar in structure to the cytoplasmic trifunctional enzyme,
except that the dehydrogenase and cyclohydrolase activities are silent due to
mutations of critical binding and catalytic residues (Christensen et al.,
404 Karen E. Christensen and Robert E. MacKenzie
2005b). Northern analysis of tissue blots (Prasannan et al., 2003) indicated
that the transcript for the MTHFD1 is highest in liver and kidney, while
that of the MTHFD1L is highest in placenta, followed by thymus, spleen,
brain, and lung, and is relatively low in liver and kidney. Sugiura et al.
(2004) found that MTHFD1L mRNA is expressed ubiquitously in normal
tissues, and of the 15 examined, found expression highest in ovary, lung,
and thymus, and lowest in white blood cells, muscle, and lymphocytes.
Using microarray gene expression profiling of normal and cancerous colon
tissue, these authors found that MTHFD1L is upregulated in colon adeno-
carcinomas. They also showed that overexpression of the enzyme in human
embryonic kidney 293 cells stimulated colony formation indicating that
expression of this gene confers a growth advantage.
V. Conclusion
A. Metabolic compartmentation of the folate enzymes
In yeast, the presence of cytoplasmic and mitochondrial isoforms of a
trifunctional dehydrogenasecyclohydrolasesynthetase support a model
wherein the mitochondria can produce formate which can be used by the
cytoplasmic enzymes for the synthesis of purines and for methylation reac-
tions. However, it is not yet clear what advantage the compartmentation of
folate metabolism provides for yeast, since deletion of the MIS1 gene has
little effect. Perhaps the mitochondrial pathway provides a useful redun-
dancy or plays a role under specific growth conditions.
In mammals, it is clear that the duplication of isoforms of the same
folate-dependent enzyme activity in both the cytoplasm and mitochondria
does support different metabolic functions. For example, the fact that
Chinese hamster ovary cells lacking mitochondrial serine hydroxymethyl-
transferase are glycine auxotrophs despite the presence of a cytosolic isoform
of the same enzyme, clearly illustrated a specific metabolic role for mito-
chondria (Chasin et al., 1974; Stover et al., 1997). The utilization of
methylenetetrahydrofolate and its interconversion with formyltetrahydro-
folate within the mitochondria enable specific roles in cellular one-carbon
metabolism. Of critical importance is the nature and distribution of the
dehydrogenase, cyclohydrolase, and synthetase activities. MTHFD2
evolved from a trifunctional dehydrogenasecyclohydrolasesynthetase
precursor with the loss of the synthetase domain (Patel et al., 2002).
Moreover, a trifunctional dehydrogenasecyclohydrolasesynthetase pre-
cursor gave rise to a monofunctional synthetase (MTHFD1L) by mutagen-
esis of critical amino acid residues essential for dehydrogenase and
cyclohydrolase activities. While all three activities are present in at least
DehydrogenaseCyclohydrolaseSynthetase 405
some mammalian mitochondria, their expression is controlled by two
different nuclear genes. Clearly one must understand why this is important.
MTHFD2 is unique in mammals because not only did it derive from a
trifunctional precursor, it also evolved to use NAD rather than NADP by an
unusual mechanism (Christensen et al., 2005a). The use of NAD instead
of NADP has been estimated to shift the equilibrium position between
methylenetetrahydrofolate and formyltetrahydrofolate in mitochondria by
as much as 200-fold in favor of formyltetrahydrofolate (Pelletier and
MacKenzie, 1995). Under conditions of rapid growth, cytoplasmic serine
is used in protein synthesis and provides methylenetetrahydrofolate required
for thymidylate synthesis and methylation reactions. To compete for serine
and ensure sufficient one-carbon units for purine synthesis, the mitochon-
drial pathway using NAD provides a thermodynamically favorable means to
convert mitochondrially generated methylenetetrahydrofolate into formate.
B. Functions of mitochondria in folate-mediated metabolism
The complement of enzymes required to produce formate from methyle-
netetrahydrofolate in the mitochondria includes both MTHFD2 and the
monofunctional MTHFD1L. When both enzymes are present in good
amounts, such as in transformed cells or in embryonic tissues, then the
pathway for mitochondrial formate production is complete. The formate
released by the mitochondria is captured by MTHFD1, and is available to
support purine synthesis. The beauty of this pathway is that whatever
cytoplasmic formyltetrahydrofolate might not be required for purine syn-
thesis can be redirected for methyl transfer reactions by MTHFD1. The
conversion of formyl- to methylenetetrahydrofolate is rate limited by the
slow conversion by the cyclohydrolase of formyl- to methenyltetrahydro-
folate (Pawelek and MacKenzie, 1998), and might contribute to the prefer-
ential access to purine synthesis for cytoplasmic formyltetrahydrofolate.
Mitochondria also use formyltetrahydrofolate to produce formylmethio-
nyl-tRNA to initiate protein synthesis. As formate equilibrates between
cytoplasmic and mitochondrial compartments, MTHFD1L can provide
formyltetrahydrofolate without relying on either mitochondrial serine or
glycine as one-carbon donors. This role is consistent with the ubiquitous
nature of its expression and can maintain mitochondrial formyltetrahydro-
folate pools to support protein synthesis in this organelle under all metabolic
conditions. The knockout of MTHFD2 (Section III.C) established the
important role of formate production during embryogenesis. One might
predict that a knockout of MTHFD1L would also be lethal because it
interrupts the pathway for mitochondrial formate production, and because
its loss might impair mitochondrial functions in cells that do not express
MTHFD2 to provide formyltetrahydrofolate.
406 Karen E. Christensen and Robert E. MacKenzie
The expression of MTHFD2 is thus the main switch to enable
mitochondria to produce formate. But what is the situation in cells that
express little, if any, of this enzyme? In most adult tissues, mitochondrial
serine hydroxymethyltransferase (SHMT) most likely equilibrates serine and
glycine in this compartment, because there is no other source of, or use for,
methylenetetrahydrofolate. However, in liver cells mitochondrial methyle-
netetrahydrofolate can also be produced via demethylation reactions
(e.g. dimethylglycine dehydrogenase) and the glycine cleavage system. In
the absence of the MTHFD2, it has been proposed that in the liver
methylenetetrahydrofolate is used with glycine by mitochondrial SHMT
to produce serine for gluconeogenesis (Christensen and MacKenzie, 2006).
Regulation of the expression of the NAD-dependent dehydrogenase
cyclohydrolase is thus a critical support to enable rapid growth when one-
carbon units for purines and glycine for heme synthesis are critical.
C. Mathematical models
Folate-mediated metabolism is a complex, nonlinear pathway, with several
sources of one-carbon units, and potential competition for these units
between methylation pathways and formyl transfer reactions. It is comfort-
able to draw these folate pathways in general diagrams (such as Fig. 14.4) but
it is becoming clearer all the time that we must be more specific with respect
to which cell, tissue, or organism or stage of development is being studied.
In recent years, the pathways have been addressed by mathematical
modeling that helps to provide insight into function, and to test hypotheses.
Reed et al. (2006) built a mathematical model for folate metabolism where
predictions matched experimental data. The model, for example, predicts
that the inverse relationship between folate and homocysteine is strongest at
very low folate concentrations, and that the DNA methylation rate is
relatively insensitive to changes in folate pool size. The model was also
used to address the conditions of vitamin B
12
deficiency and polymorphisms
in methylenetetrahydrofolate reductase and the effects on purine and thy-
midine synthesis. More recently, the model has been refined to address the
compartmentalization of folate-mediated metabolism (Nijhout et al., 2006).
It predicts the critical role of MTHFD2 that enables mitochondria to
produce formate to support purine synthesis in transformed cells and
embryonic tissues, while in adult hepatic cells the mitochondria produce
serine from glycine that can be used for gluconeogenesis. The development
of models like this one will assist in the understanding of folate metabolism
when comparing different types of cells and tissues, and help predict the
effects of genetic or chemical alterations of these pathways. They hold
promise as useful tools in predicting the effects of inhibitors of various
targets in the pathway, specific for cell type, including tumor cells.
DehydrogenaseCyclohydrolaseSynthetase 407
REFERENCES
Allaire, M., Li, Y., MacKenzie, R. E., and Cygler, M. (1998). The 3-D structure of a folate-
dependent dehydrogenase/cyclohydrolase bifunctional enzyme at 1.5 A
resolution.
Structure 6, 173182.
Appling, D. R. (1991). Compartmentation of folate-mediated one-carbon metabolism in
eukaryotes. FASEB J. 5, 26452651.
Barlowe, C. K., and Appling, D. R. (1988). In vitro evidence for the involvement of
mitochondrial folate metabolism in the supply of cytoplasmic one-carbon units. Biofactors
1, 171176.
Belanger, C., and MacKenzie, R. E. (1989). Isolation and characterization of cDNA clones
encoding the murine NAD-dependent methylenetetrahydrofolate dehydrogenase-
methenyltetrahydrofolate cyclohydrolase. J. Biol. Chem. 264, 48374843.
Chasin, L. A., Feldman, A., Konstam, M., and Urlaub, G. (1974). Reversion of a Chinese
hamster cell auxotrophic mutant. Proc. Natl. Acad. Sci. USA 71, 718722.
Christensen, K. E., and MacKenzie, R. E. (2006). Mitochondrial one-carbon metabolism is
adapted to the specific needs of yeast, plants and animals. BioEssays 28, 595605.
Christensen, K. E., Mirza, I. A., Berghuis, A. M., and MacKenzie, R. E. (2005a). Magnesium
and phosphate ions enable NAD binding to methylenetetrahydrofolate dehydrogenase-
methenyltetrahydrofolate cyclohydrolase. J. Biol. Chem. 280, 3431634323.
Christensen, K. E., Patel, H., Kuzmanov, U., Mejia, N. R., and MacKenzie, R. E. (2005b).
Disruption of the Mthfd1 gene reveals a monofunctional 10-formyltetrahydrofolate
synthetase in mammalian mitochondria. J. Biol. Chem. 280, 75977602.
Cohen, L., and MacKenzie, R. E. (1978). Methylenetetrahydrofolate dehydrogenase-
methenyltetrahydrofolate cyclohydrolase-formyltetrahydrofolate synthetase from porcine
liver. Interaction between the dehydrogenase and cyclohydrolase activities of the
multifunctional enzyme. Biochim. Biophys. Acta 522, 311317.
DeLano, W. L. (2004). The PyMOL Molecular Graphics System. DeLano Scientific
LLC, San Carlos, CA, USA. http://www.pymol.org.
Di Pietro, E., Sirois, J., Tremblay, M. L., and MacKenzie, R. E. (2002). Mitochondrial
NAD-dependent methylenetetrahydrofolate dehydrogenase-methenyltetrahydrofolate
cyclohydrolase is essential for embryonic development. Mol. Cell. Biol. 22, 41584166.
Di Pietro, E., Wang, X. L., and MacKenzie, R. E. (2004). The expression of mitochondrial
methylenetetrahydrofolate dehydrogenase-cyclohydrolase supports a role in rapid cell
growth. Biochim. Biophys. Acta 1674, 7884.
Hum, D. W., and MacKenzie, R. E. (1991). Expression of the active domains of the human
folate-dependent trifunctional enzyme in E. coli. Protein Eng. 4, 493500.
Jones, E. W. (1977). Bipartite structure of the ade3 locus of Saccharomyces cerevisiae. Genetics
85, 209223.
MacKenzie, R. E. (1984). Interconversion of substituted tetrahydrofolates. In Folates and
Pterins: Chemistry and Biochemistry of Folates (R. Blakely and S. Benkovic, eds.), Vol. 1,
pp. 255306. John Wiley and Sons, NewYork.
Mejia, N. R., and MacKenzie, R. E. (1985). NAD-dependent methylenetetrahydrofolate
dehydrogenase is expressed by immortal cells. J. Biol. Chem. 260, 1461614620.
Mejia, N. R., and MacKenzie, R. E. (1988). NAD-dependent methylenetetrahydrofolate
dehydrogenase-methenyltetrahydrofolate cyclohydrolase in transformed cells is a mito-
chondrial enzyme. Biochem. Biophys. Res. Commun. 155, 16.
Mejia, N. R., Rios-Orlandi, E. M., and MacKenzie, R. E. (1986). NAD-dependent
methylenetetrahydrofolate dehydrogenase-methenyltetrahydrofolate cyclohydrolase
from ascites tumor cells. Purification and properties. J. Biol. Chem. 261, 95099513.
408 Karen E. Christensen and Robert E. MacKenzie
Nijhout, H. F., Reed, M. C., Lam, S. L., Shane, B., Gregory, J. F., III, and Ulrich, C. M.
(2006). In silico experimentation with a model of hepatic mitochondrial folate metabo-
lism. Theor. Biol. Med. Model. 3, 40; doi: 10.1186/1742-4682-3-40.
Patel, H., Christensen, K. E., Mejia, N., and MacKenzie, R. E. (2002). Mammalian
mitochondrial methylenetetrahydrofolate dehydrogenase-cyclohydrolase derived from
a trifunctional methylenetetrahydrofolate dehydrogenase-cyclohydrolase-synthetase.
Arch. Biochem. Biophys. 403, 145148.
Patel, H., Di Pietro, E., and MacKenzie, R. E. (2003). Mammalian fibroblasts lacking mito-
chondrial NAD
-dependent methylenetetrahydrofolate
dehydrogenase-cyclohydrolase: Detection of the mRNA in normal murine tissues and
transcriptional regulation of the gene in cell lines. Biochim. Biophys. Acta 1171, 281287.
Prasannan, P., Pike, S., Peng, K., Shane, B., and Appling, D. R. (2003). Human mitochon-
drial C1-tetrahydrofolate synthase: Gene structure, tissue distribution of the mRNA, and
immunolocalization in Chinese hamster ovary cells. J. Biol. Chem. 278, 4317843187.
Reed, M. C., Nijhout, H. F., Neuhouser, M. L., Gregory, J. F., III, Shane, B., James, S. J.,
Boynton, A., and Ulrich, C. M. (2006). A mathematical model gives insights into
nutritional and genetic aspects of folate-mediated one-carbon metabolism. J. Nutr. 136,
26532662.
Rios-Orlandi, E. M., and MacKenzie, R. E. (1988). The activities of the NAD-dependent
methylenetetrahydrofolate dehydrogenase-cyclohydrolase from ascites tumor cells are
kinetically independent. J. Biol. Chem. 263, 46624667.
Schmidt, A., Wu, H., MacKenzie, R. E., Chen, V. J., Bewly, J. R., Ray, J. E., Toth, J. E.,
and Cygler, M. (2000). Structures of three inhibitor complexes provide insight into the
reaction mechanism of the human methylenetetrahydrofolate dehydrogenase/cyclohy-
drolase. Biochemistry 39, 63256335.
Scrimgeour, K. G., and Huennekens, F. M. (1960). Occurrence of a DPN-linked N
5
, N
10
-
methylenetetrahydrofolate dehydrogenase in Ehrlich ascites tumor cells. Biochem.
Biophys. Res. Commun. 2, 230233.
DehydrogenaseCyclohydrolaseSynthetase 409
Shannon, K. W., and Rabinowitz, J. C. (1986). Purification and characterization of a
mitochondrial isozyme of C1-tetrahydrofolate synthase from Saccharomyces cerevisiae.
J. Biol. Chem. 261, 1226612271.
Shannon, K. W., and Rabinowitz, J. C. (1988). Isolation and characterization of the
Saccharomyces cerevisiae MIS1 gene encoding mitochondrial C1-tetrahydrofolate synthase.
J. Biol. Chem. 263, 77177725.
Smith, D. D. S., and MacKenzie, R. E. (1983). Methylenetetrahydrofolate dehydrogenase-
methenyltetrahydrofolate cyclohydrolase-formyltetrahydrofolate synthetase fromporcine
liver: Evidence to support a common dehydrogenase-cyclohydrolase site. Can. J. Biochem.
Cell Biol. 61, 11661171.
Smith, G. K., Banks, S. D., Monaco, T. J., Rigual, R., Duch, D. S., Mullin, R. J., and
Huber, B. E. (1990). Activity of an NAD-dependent 5,10-methylenetetrahydrofolate
dehydrogenase in normal tissue, neoplastic cells, and oncogene-transformed cells. Arch.
Biochem. Biophys. 283, 367371.
Staben, C., and Rabinowitz, J. C. (1986). Nucleotide sequence of the Saccharomyces cerevisiae
ADE3 gene encoding C1-tetrahydrofolate synthase. J. Biol. Chem. 261, 46294637.
Stover, P. J., Chen, L. H., Suh, J. R., Stover, D. M., Keyomarsi, K., and Shane, B. (1997).
Molecular cloning, characterization, and regulation of the human mitochondrial serine
hydroxymethyltransferase gene. J. Biol. Chem. 272, 18421848.
Sugiura, T., Nagano, Y., Inoue, T., and Hirotani, K. (2004). A novel mitochondrial C
1
-
tetrahydrofolate synthetase is upregulated in human colon adenocarcinoma. Biochem.
Biophys. Res. Commun. 315, 204211.
Sundararajan, S., and MacKenzie, R. E. (2002). Residues involved in the mechanism of the
bifunctional methylenetetrahydrofolate dehydrogenase-cyclohydrolase. J. Biol. Chem.
277, 1870318709.
Tan, L. U., Drury, E. J., and MacKenzie, R. E. (1977). Methylenetetrahydrofolate
dehydrogenase-methenyltetrahydrofolate cyclohydrolase-formyltetrahydrofolate synthe-
tase. A multifunctional enzyme from porcine liver. J. Biol. Chem. 252, 11171122.
Thigpen, A. E., West, M. G., and Appling, D. R. (1990). Rat C
1
-tetrahydrofolate synthase.
J. Biol. Chem. 256, 79077913.
Walkup, A. S., and Appling, D. R. (2005). Enzymatic characterization of human mitochon-
drial C-1-tetrahydrofolate synthase. Arch. Biochem. Biophys. 442, 196205.
Woeller, C. F., Anderson, D. D., Szebenyi, D. M. E., and Stover, P. J. (2007). Evidence for
small ubiquitin-like modifier-dependent nuclear import of the thymidylate biosynthesis
pathway. J. Biol. Chem. 282, 1762317631.
Yang, X. M., and MacKenzie, R. E. (1993). NAD dependent methylenetetrahydrofolate
dehydrogenase-methenyltetrahydrofolate cyclohydrolase is the mammalian homolog of
the mitochondrial enzyme encoded by the yeast MIS1 gene. Biochemistry 32,
1111811123.
410 Karen E. Christensen and Robert E. MacKenzie
C H A P T E R F I F T E E N
The Structure and Mechanism of
6-Hydroxymethyl-7,8-Dihydropterin
Pyrophosphokinase
Jeremy P. Derrick*
Contents
I. Introduction 412
II. Structural and Mechanistic Studies on EcoHPPK 413
A. Identification of the HPPK fold and substrate binding 413
B. Mechanism of catalysis 419
C. Inhibitors 420
III. Structures of HPPKs from Other Organisms 421
IV. Kinetics 424
V. Relationship of HPPK to Other Pyrophosphoryl Transfer Enzymes 426
VI. Concluding Remarks 429
References 430
Abstract
6-Hydroxymethyl-7,8-dihydropterin pyrophosphokinase (HPPK) catalyses the
transfer of pyrophosphate from ATP to 6-hydroxymethyl-7,8-dihydropterin (HMDP),
and is an essential enzyme in the biosynthesis of folic acid. It is also a potential
target for antimicrobial drugs. HPPK from Escherichia coli, which has been the
most intensively investigated, is a monomeric protein with a molecular mass of
about 18,000. Structures of the enzyme, determined by X-ray crystallography
and NMR, have shown that it adopts an a/b fold with a substrate-binding cleft on
the surface. Three loop regions surround the enzyme active site and form
intimate contacts with the substrates. The enzyme has a fixed order of substrate
binding, with ATP binding first, followed by HMDP. Binding of ATP causes a shift
in the conformations of the loop regions, which completes formation of the
HMDP-binding site. Two magnesium ions bind within the active site, bridging
between the phosphate groups in ATP and the enzyme. Both ions appear to play
an integral role in ATP recognition and stabilization of the transition state of the
reaction. Ligand binding and kinetic studies have shown that the overall rate of
Vitamins and Hormones, Volume 79
#
2008 Elsevier Inc.
ISSN 0083-6729, DOI: 10.1016/S0083-6729(08)00415-9 All rights reserved.
* Faculty of Life Sciences, Manchester Interdisciplinary Biocentre, University of Manchester, Manchester,
United Kingdom
411
the reaction is not limited by the rate of substrate transformation into products
on the enzyme, which is relatively fast, but is more likely caused by a slow step
associated with product release. These fundamental studies open up the potential
for exploitation through the design of specific HPPK inhibitors. 2008 Elsevier Inc.
I. Introduction
6-Hydroxymethyl-7,8-dihydropterin pyrophosphokinase (HPPK; EC
2.7.6.3) catalyses an essential step in the biosynthesis of folic acid. The
enzyme was first identified in 1969 (Richey and Brown, 1969) but it is in
the last 8 years that most of our current knowledge of its structure and
mechanism has been acquired. Folic acid is a dietary requirement for man,
but bacteria, plants, and parasites have the ability to synthesize folic acid
de novo. As a result, HPPK, along with several of the enzymes from the folate
pathway, are not found in man, rendering them attractive potential targets
for the development of antimicrobial agents. Indeed, the enzyme in the
folate pathway after HPPK, dihydropteroate synthase (DHPS), has been
known for many years to be the target for sulfonamide drugs (Bermingham
and Derrick, 2002). The possibility that knowledge of the structure and
mechanism of HPPK could be deployed to direct the development of novel
antimicrobials is a strong justification for research into the enzyme.
The section of the folic acid biosynthesis pathway to which HPPK
belongs is shown in Fig. 15.1. Although it is not shown in Fig. 15.1, the
pathway begins with formation of the pterin ring from GTP, via a complex
O
OH
CH
2
OH DHNA
HO
N
N
HN
H
2
N N
H
7,8-dihydroneopterin Glycoladehyde
O
CH
2
OH
(HMDP)
ATP
AMP
HPPK
N
N
HN
H
2
N N
H
6-hydroxymethyl-
7,8-dihydropterin
O
CH
2
O
N P P
P P
N
HN
NH
2
H
2
N N
H
pABA
DHPS
6-hydroxymethyl-
7,8-dihydropterin
pyrophoshate
(DHPPP)
O
CH
2
NH N
CO
2
N
HN
H
2
N
CO
2
N
H
7,8-dihydropteroate
Figure 15.1 Three steps from the folic acid biosynthesis pathway.
412 Jeremy P. Derrick
reaction catalyzed by GTP cyclohydrolase I (Auerbach et al., 2000; Bracher
et al., 2001; Nar et al., 1995, 1996). An intermediate in the pathway, 7,8-
dihydroneopterin, is a substrate for dihydroneopterin aldolase (DHNA),
which releases glycolaldehyde to form 6-hydroxymethyl-7,8-dihydropterin
(HMDP). HPPK then catalyses the transfer of pyrophosphate from ATP to
HMDP, to form 6-hydroxymethyl-7,8-dihydropterin pyrophosphate
(DHPPP) and AMP. DHPPP acts as a substrate for DHPS, which uses
para-aminobenzoic acid (pABA) to form dihydropteroate. DHPS can be
thought of acting as a link between the folate pathway and the chorismate
pathway, which synthesizes pABA (Dosselaere and Vanderleyden, 2001).
The final steps in the pathway then form dihydrofolate and tetrahydrofolate
[not shown in Fig. 15.1a more complete review of the folate biosynthesis
pathway can be found in Bermingham and Derrick (2002)].
The evidence for the pathway in Fig. 15.1 comes from the pioneering
experiments of Gene Brown and colleagues. In 1961, Brown et al. estab-
lished that cell free extracts from Escherichia coli were able to convert
HMDP, ATP/Mg
2
, and pABA to dihydropteroate (Brown et al., 1961).
Further work provided evidence that DHPPP was an intermediate in the
synthetic pathway (Weisman and Brown, 1964). Richey and Brown then
demonstrated that two distinct enzymes were responsible for dihydroptero-
ate formation and partially separated them by chromatography (Richey and
Brown, 1969). At this stage it was clear that there was a distinct enzyme
which was responsible for the transfer of pyrophosphate from ATP to
HMDP. In 1990, Lopez et al. reported the DNA sequencing, partial purifi-
cation, and overexpression of HPPK from Streptococcus pneumoniae (Lopez
et al., 1990). E. coli HPPK (EcoHPPK), which was to form the basis for
much structural and kinetic investigation, was purified to homogeneity by
Talarico et al. (1991), a significant achievement given that the enzyme
constituted only about 0.01% of soluble cell protein. This feat permitted
the subsequent cloning, sequencing, and overexpression of the E. coli
enzyme (Talarico et al., 1992). Since then, well over 600 separate HPPK
sequences have been determined from different organisms (as listed by Pfam
(Bateman et al., 2004)). The purpose of this review is to focus specifically on
recent structural and kinetic studies, which have ensured that HPPK is
currently the best understood pyrophosphoryl transfer enzyme to date.
II. Structural and Mechanistic Studies on
EcoHPPK
A. Identification of the HPPK fold and substrate binding
The first 3-D structures of HPPK were determined independently for
enzymes derived from E. coli (Stammers et al., 1999; Xiao et al., 1999) and
Haemophilus influenzae (Hennig et al., 1999). The enzyme adopts an ab
Structure and Mechanism of HPPK 413
sandwich foldbroadly similar to ferridoxinbut the precise arrangement
of secondary structural elements is, so far, unique to HPPK. The fold
comprises six b-stands and four a-helices in the order b1-a1-b2-b3-a2-
b4-b5-b6-a3-a4, such that the six strands form a single antiparallel b-sheet,
with the helices arranged on either side (Fig. 15.2A). The active site has been
identified by cocrystallization with a variety of substrates and inhibitors
details of some of the structures determined are summarized in Table 15.1.
The substrates bind within a valley on the enzyme surface, 25 A
long and
10 A
deep; the site lies on one face of the central b-sheet and is bounded at
opposite ends by the a2 and a3 helices. Amajor feature of ligand recognition
in HPPK is mediated by three loop regions, between b1 and a1 (loop L1),
b2 and b3 (loop L2), and a2 and b4 (loop L3). The dynamic behavior of these
loop regions plays a major role in substrate recognition and catalysis.
Several investigators have useda varietyof substrates andsubstrateanalogues
tostudy the residues withinHPPKthat are involvedinrecognitionof ATPand
HMDP. Hennig et al. (1999) showed by ultracentrifugation that the substrate
analog 6-hydroxymethyl-7,7-dimethyl-7,8-dihydropterin (HMMDP) bound
to H. influenzae HPPK (HiHPPK). This ligand was found to be degraded to
6-hydroxy-7,7-dimethyl-7,8-dihydropterin on prolonged incubation with
the enzyme, however, and the 2.1 A
) Ligands References
Escherichia coli 1HKA XRD 1.50 None Xiao et al., 1999
Escherichia coli 2F65 NMR AMPCPP Li et al., 2006
Escherichia coli 2F63 NMR AMPCPP and
6-hydroxymethyl-7,7-
dimethylpterin
Li et al., 2006
Escherichia coli 1EQM XRD 1.50 ADP Xiao et al., 2001
Escherichia coli 1RB0 XRD 1.35 DHPPP Blaszczyk et al., 2004b
Escherichia coli 1Q0N XRD 1.25 AMPCPP and HMDP Blaszczyk et al., 2000
Escherichia coli 1RAO XRD 1.56 AMP and DHPPP Blaszczyk et al., 2004b
Escherichia coli 1DY3 XRD 2.0 ATP and 6-hydroxymethyl-7-
methyl-7-phenethyl-7,8-
dihydropterin
Stammers et al., 1999
Escherichia coli 1EX8 XRD 1.85 6-(Adenosine tetraphosphate-
methyl)-7,8-dihydropterin
Shi et al., 2001
Haemophilus
inuenzae
1CBK XRD 2.1 6-Hydroxymethyl-7,7-
dimethyl-7,8-dihydropterin
Hennig et al., 1999
Saccharomyces
cerevisiae
2BMB XRD 2.3 6-Hydroxymethylpterin
monophosphate
Lawrence et al., 2005
Streptococcus
pneumoniae
2CG8 XRD 2.9 None Garcon et al., 2006
The crystal structures of the ternary complexes of EcoHPPK established
a key role for the two Mg
2
ions which are bound within the active site.
Catalysis is known to be dependent on Mg
2
, and the ion also has an effect
on the binding affinity of the enzyme for ATP (Shi et al., 2000). Both ions
are 6-coordinated, one bridging the a- and b-phosphates of AMPCPP/
ATP, and the other bridging the b- and g-phosphates (Fig. 15.3A). Addi-
tional coordination is provided from the side chains of two absolutely
conserved Asp residues, D95 and D97, and the hydroxyl group of
HMDP. Other residues involved in binding ATP are E77, R92, H115,
and R121. Once ATP has bound, formation of the HMDP-binding pocket
is completed. It is highly specific for recognition of the pterin ring: all the
potential hydrogen bond donors and acceptors on the pterin ring in posi-
tions 1, 2, 3, 4, and 8 are saturated through hydrogen bonds to residues
Figure 15.3 Substrate recognition by E. coli HPPK (EcoHPPK) (A) Interaction
between a,b-methyleneadenosine triphosphate (AMPCPP), D95, D97, and the two
Mg
2
ions within the active site. (B) Residues involved in binding 6-hydroxymethyl-
7,8-dihydropterin (HMDP). Selected residues are indicated, superimposed on a blue
ribbon plot of the EcoHPPK backbone. The AMPCPP molecule has been omitted for
clarity. Coordinates were taken from PDB code 1Q0N for both diagrams.
Structure and Mechanism of HPPK 417
within the binding site. Furthermore, the pterin ring is effectively sand-
wiched on both sides by two aromatic residues, Y53 and F123. An illustra-
tion of the principal interacting residues is shown in Fig. 15.3B. The highly
specific nature of the recognition of the pterin ring explains the high affinity
of the enzymeAMPCPP binary complex for HMDP (Bermingham et al.,
2000; Li et al., 2002). It also has important ramifications for the design of
inhibitors against the enzyme (see following section).
Two product complex structures with EcoHPPK have been reported,
with DHPPP alone and AMPDHPPP (Blaszczyk et al., 2004b). The
authors used these to complete the series of enzyme-substrate and enzyme-
product crystal structures that describe the HPPK reaction pathway. They
observed that there is a significant degree of disorder in the phosphate groups
in AMP and DHPPP, requiring the modeling of two alternate conforma-
tions. They also reported that the AMPDHPPP ternary complex had an
almost identical conformation to an NMR structure for the EcoHPPK
AMPCPP binary complex, with loop L3 displaced far from the active
site. By contrast, loop L3 is closed over the active site in the EcoHPPK
AMPCPPHMDP ternary complex. The authors interpreted this series of
structures as indicative of extensive movement of loop L3 during substrate
binding, catalysis, and product release. The loop is displaced outwards on
ATP binding, inwards after formation of the AMPCPPHMDP ternary
complex and back outwards again after formation of the EcoHPPK
AMPDHPPP product complex.
The conclusions discussed above concerning structural transitions in
HPPK during the reaction cycle were based solely on crystal structures of
the various intermediate structural states. It is well established that the
constraints of packing within a crystalline lattice can fix loop regions in
particular conformational states, which may not be accurately representative
of the situation in solution. In the case of HPPK, more recent work by
NMR (Li et al., 2006) and computational studies (Keskin et al., 2002; Yang
et al., 2005) has indicated that the loop regions in the enzyme are accessible
to a wider range of conformational states than is apparent from the crystal
structure, particularly of the apoenzyme form.
As a consequence of the small size and monomeric state of EcoHPPK,
solution state NMR has made a valuable contribution to structural studies
on the enzyme (Garcon et al., 2004; Li et al., 2006; Xiao et al., 2001). The
3-D structure of EcoHPPK in complex with the ATP analog b,g-
methyleneadenosine triphosphate (AMPPCP) provided a verification,
independent from crystallography, of structural transitions which occur on
the binding of ATP (Xiao et al., 2001). More recently, two further struc-
tures of a binary complex with AMPCPP and a ternary complex with
AMPCPP and 6-hydroxy-7,7-dimethyl-pterin, have provided novel
insights into the nature of the conformational changes between different
ligand-bound states of the enzyme (Li et al., 2006). Both the apoenzyme and
418 Jeremy P. Derrick
the binary complexes with ATP analogs exhibit a higher degree of confor-
mational flexibility, particularly within loop regions L1L3, than the ternary
complex. This is manifest in weak or absent NH cross peaks for residues
within loops L1 and L3 for the AMPCPP binary complex. Most signals
from residues in the ternary complex, by contrast, are well defined, includ-
ing all those originating from loops L1 to L3. These observations suggest a
tightening-up of the structure on binding of the HMDP analog. The
results provide evidence for multiple structural conformations in the apo-
enzyme and binary complexes in solution, requiring a revision of the
classical induced fit model of substrate binding, as applied to EcoHPPK,
to a more complex population-based model (Li et al., 2006).
The conclusions from NMR work have been bourne out by computa-
tional studies of the dynamics of the EcoHPPK structure: Keskin et al.
(2002) concluded that the core regions of secondary structure within
HPPK, which comprise most of the residues conserved in sequence, were
rigid but that the L2 and L3 loops exhibited a much greater degree of
motion. They also provided independent evidence of the relative restriction
in mobility of the structure on binding of substrates. Yang et al. (2005)
applying a different methodology (essential dynamics analysis) reached
similar conclusions and suggested that the open conformation of L3,
which had been identified in the crystal structure complex with ADP
(Xiao et al., 2001), was also accessible to the loop in solution in the apoen-
zyme form. The current consensus is that a selected fit model pertains to
the initial binding events in the HPPKreaction cycle: the apoenzyme adopts a
wide range of conformational states, principally driven by diverse structures
of L2 and L3. Following binding of ATP, a reorganization of the L3
structure promotes formation of the HMDP-binding site. It would be
interesting to pursue the study of other structural states of the EcoHPPK
reaction cycle, particularly the product complexes, given that the steps
associated with product departure are apparently responsible for limiting
the overall rate of catalysis (discussed further in Section IV, below).
B. Mechanism of catalysis
The current model for the mechanism of the pyrophosphoryl transfer
reaction is based primarily on crystallographic data, particularly the struc-
ture of the nonproductive ternary complex EcoHPPKAMPCPPHMDP
(Blaszczyk et al., 2000). The authors proposed an in-line single displacement
mechanism for nucleophilic attack of the b-phosphorus in ATP by the
hydroxyl oxygen in HMDP. It was suggested that the reaction has some
associative character (shown schematically in Fig. 15.4): an associative
mechanism would require that the transition state contains a pentacoordi-
nate geometry at the b-phosphorus atom in the transition state (S
N
2-like).
By contrast, a dissociative transition state would have a trigonal planar
Structure and Mechanism of HPPK 419
geometry at this atom (S
N
1-like) (Matte et al., 1998). In the case of HPPK,
the value of the EcoHPPKAMPCPPHMDP structure in distinguishing
between these possibilities lies in its degree of similarity with the real
EcoHPPKATPHMDP transition state.
A number of roles in substrate binding and catalysis have been proposed
by Blaszczyk et al. (2000) for the two Mg
2
ions that are bound within the
HPPK active site. It seems likely that the Mg
2
ions play a role in ATP
binding and the associated conformational change which promotes the
binding of HMDP. The ions could play an important part in orientating
the hydroxyl oxygen of HMDP relative to the b-phosphorus and the
bridging oxygen atom between the a- and b-phosphates to ensure optimal
geometry for the reaction. The Mg
2
ions could also activate the b-phos-
phorus for nucleophilic attack, and contribute to stabilization of the negative
charge on the transition state. Again, these proposals are largely based on the
crystal structure and require further investigation by other techniques.
C. Inhibitors
Relatively few specific inhibitors of HPPK have been reported in the
literature. Wood and colleagues described the synthesis of disubstituted
pterin analogs with two substituents at the 7-position on the ring, which
were shown to inhibit HPPK activity (Wood, 1975). The use of these
inhibitors in structural studies on E. coli (Li et al., 2006; Stammers et al.,
1999) and HiHPPK (Hennig et al., 1999) has already been discussed.
A comparison of the structures of the EcoHPPKATP6-hydroxymethyl-
7-methyl-7-phenethyl-7,8-dihydropterin (PDB 1DY3) and EcoHPPK
AMPCPPHMDP (PDB 1Q0N) ternary complexes shows that W89 in
the latter structure would cause a steric clash with the phenethyl substituent.
The side chains of R82 and R92, known to be essential for catalysis (Li et al.,
2003), adopt different rotamers in the two structures. Furthermore, the
conformation of L3 differs between the two structures. These slight changes
Figure 15.4 Schematic diagram of the transition state for the 6-hydroxymethyl-7,8-
dihydropterin pyrophosphokinase (HPPK) reaction. Adapted from Blaszczyk et al.
(2000).
420 Jeremy P. Derrick
to the geometry of the active site could explain why 6-hydroxymethyl-7-
methyl-7-phenethyl-7,8-dihydropterin does not appear to be a substrate for
the reaction.
An alternative approach to the inhibition of EcoHPPK was developed
by Shi et al. (2001) who experimented with three different bisubstrate
analogs: 6-hydroxymethylpterin was linked by 2, 3, or 4 phosphate groups
to adenosine. Optimal inhibition was obtained with the tetraphosphate
inhibitor, which bound to the enzyme with submicromolar affinity.
A crystal structure of the inhibitor bound to EcoHPPK showed that the
ligand did indeed form a bridge between the pterin and ATP-binding sites.
The work illustrated the feasibility of designing HPPK inhibitors based on
bisubstrate analogs.
III. Structures of HPPKs from Other Organisms
The crystal structures of several other HPPK enzymes have been
reported from different organisms. The structure of HPPK from HiHPPK
was reported at about the same time as the first EcoHPPK crystal structure,
and is very similar in many respects (Hennig et al., 1999). It was determined
in complex with the inhibitor 6-hydroxy-7,7-dimethyl-7,8-dihydropterin
but, interestingly, without ATP bound: it appears that the pterin-based
inhibitor is capable of binding to HiHPPK in the absence of ATP. HiHPPK
is a monomer in solution although the crystallographic asymmetric unit
contains a dimer; the formation of the dimer is a feature of packing
within the crystal form, and unlikely to have any physiological significance.
The conformation of the loops in the HiHPPK structure is closest to the
EcoHPPKAMPCPPHMDP ternary complex (Fig. 15.5A) and it is likely
that HiHPPK passes through a similar closed state in its reaction cycle, in
a similar fashion to EcoHPPK.
In parasites, plants, and some bacteria, the HPPK polypeptide is fused to
the preceding and/or following enzymes in the folate pathway, generating a
multienzyme complex. In the malaria parasite Plasmodium falciparum, for
example, HPPK is fused to DHPS, the following enzyme in the pathway
and the target for the sulfonamide group of drugs (Brooks et al., 1994). HPPK
DHPS fusions are also found in plants (Rebeille et al., 1997). Storozhenko et al.
give details of HPPKDHPS sequences from29 different organisms, covering
plants, fungi, and some eubacteria (Storozhenko et al., 2007). By contrast,
HPPK from the respiratory pathogen S. pneumoniae is joined to the preceding
enzyme, DHNA (Lopez and Lacks, 1993). In Pneumocystis carinii, a trifunc-
tional DHNAHPPKDHPS complex has been found (Volpe et al., 1993).
The reason why the folate enzymes from such a diverse range of organisms
have a tendency to form multienzyme complexes is unclear. It is possible that
Structure and Mechanism of HPPK 421
some advantage may be gained by substrate channeling: this is a phenomenon
where the product of one enzyme is handed on to the next enzyme in the
pathway, without equilibrating with the bulk phase. It is well established that
Figure 15.5 Comparison of 6-hydroxymethyl-7,8-dihydropterin pyrophosphokinase
(HPPK) structures. (A) Structural alignment of H. influenzae HPPK (HiHPPK) (PDB
1CBK) in red with E. coli HPPK (EcoHPPK) (PDB 1Q0N) in green. The inhibitor
6-hydroxy-7,7-dimethyl-7,8-dihydropterin is superimposed. (B) Structural alignment
of the HPPK domain from S. cerevisiae (PDB 2BMB) in red with EcoHPPK (PDB
1Q0N) in green. The ligand 6-hydroxymethylpterin monophosphate, which was
cocrystallized with the Saccharomyces enzyme, is superimposed. Loops L2 and L3 are
indicated. (C) Assembly of S. cerevisiae HPPKdihydropteroate synthase (ScHPPK
DHPS). The DHPS domains are in red, HPPK domains in green and the linker peptide
(between DHPS and HPPK) is in blue. The 6-hydroxymethylpterin monophosphate
ligands show the positions of the active sites. (D) S. pneumoniae dihydroneopterin
aldolase (DHNA)HPPK (PDB 2CG8). The locations of the HPPK and DHPS domains
are indicated.
422 Jeremy P. Derrick
this can occur in enzymes from some biosynthetic pathways (Huang et al.,
2001). However, there is currently no evidence for channeling of substrates to,
or products from HPPK.
In the yeast Saccharomyces cerevisiae, HPPK is found as part of a trifunc-
tional DHNAHPPKDHPS polypeptide; in order to investigate the
assembly of this complex, Lawrence et al. (2005) determined the crystal
structure of the HPPKDHPS fragment to 2.3 A
resolution. A structural
alignment of the S. cerevisiae HPPK (ScHPPK) component with EcoHPPK
shows good agreement with the location of most secondary structures,
although electron density for the equivalent of loop 3 (L3) is missing from
the Saccharomyces structure (Fig. 15.5B). Interestingly, L2 adopts a confor-
mation in the ScHPPK structure which is similar to its equivalent in the
EcoHPPKAMPCPPHMDP ternary complex. The main point of diver-
gence between the two HPPK structures lies in an additional five residues in
ScHPPK, between b4 and a3. Crystal structures of DHPS show that the
enzyme adopts a TIM barrel-type fold and is dimeric (Achari et al., 1997;
Babaoglu et al., 2004; Baca et al., 2000; Hampele et al., 1997). The associa-
tion of the two DHPS monomers in the ScHPPKDHPS complex is the
dominant feature of the structure (Fig. 15.5C). The two HPPK domains are
kept apart, and do not make contact. The structure was determined with the
inhibitor 6-hydroxymethylpterin monophosphate (PMM) bound in both
the HPPK and the DHPS active sites; the two active sites are well separated
in space, with a distance of about 34 A